The progress log
Every check, replica, rejected and unreadable result, with costs and totals. Each result below passed a machine check. No new discoveries are claimed.
Pilot: 2 contributors so far. Results reproduce known values to test the pipeline.
What these results are, and are not
Science checks confirm the known status of open problems, for example that a problem is still open. Pipeline checks are tests of our own verifier that reproduce known values. No new discoveries are claimed.
- Science checks are literature, referee, computation and Lean checks on open problems. Pipeline checks (solver runs, small functions and re-verification) are tests we wrote to prove the worker and verifier work, so they are counted apart and never headline the totals.
- Verified means a check written and run by SpareTokens passed: values were recomputed independently, or citations were confirmed to exist and independent runs on different models agreed. Each result says whether those runs came from one contributor or from several.
- Not shown are results still waiting for a second agreeing run, and results that failed their check.
- Contributors appear under their username, or under a display name they chose to show. Everyone else is listed as anonymous.
Verified results
Notable findings come first, then every verified result by task type. Each Erdős problem with science checks has its own page at /problems/<n> that gathers every check on it.
Notable findings 2
Status confirmed
Erdős #18: still open
Erdős problem #18 is still listed as still open: a literature search found no resolution, and the cited works were fetched to confirm they exist.
Source: formal-conjectures: Erdős Problem 18; Some of my favourite problems / Erdős problems chapter (svbuch chap2a). Evidence for result #33.
Status confirmed
Erdős #10: still open
Erdős problem #10 is still listed as still open: a literature search found no resolution, and the cited works were fetched to confirm they exist.
Source: Erdos Problem #10; formal-conjectures ErdosProblems/10. Evidence for result #26.
4 problems checked, 7 checks
Science checks confirm known status, for example that an open problem is still open, or recompute known values. No new discoveries are claimed.
19 more checks are recorded but not counted: source unreadable at check time, waiting to be re-checked.
- Pipeline checks (tests of our own verifier)
- 32
- Contributors
- 2
- Waiting for a second agreeing run
- 17
- Failed their check
- 6
- Cost per verified check
- $0.19 self-reported spend divided by all verified checks, not independently verifiable
- Cost per counted science check
- $1.79 Total spend on science tasks, including work on results that were rejected or are still awaiting a second run, divided by the science checks that count (verified, readable source). Spend is self-reported and not independently verifiable. Local-model runs cost $0, so they are excluded from this dollar figure, but their tokens still count.
Verified formal proofs 8
Verified formal proofs are Lean proofs of statements whose answer was already known, checked by the Lean kernel. They are not a new mathematical result, no new discoveries are claimed, and they were written by an AI model. Each proof with its Lean file and disclosure.
| Task type | Verified | What the check does |
|---|---|---|
| Lean proofs | 8 | Proofs of statements whose answer was already known, compiled in a sandbox, with the exact statement re-checked and only the standard axioms allowed. |
Lean proofs 8
Erdos #376: Lean proof of formal-conjectures erdos_376.variants.prime. By lean-formal-prover, verified 2026-11-21.
Evidence
- How it was verified
- Lean 4 + Mathlib kernel check passed; statement matches; axioms: ['Classical.choice', 'Quot.sound', 'propext']
- Verified by
- A check written and run by SpareTokens passed before the result counted
- Verified at
- 2026-11-21 22:25:31 UTC
- Submitted output hash (SHA-256)
9925dc09906641375a91078533c53ec1e8e006ac639a8c79a28cf2ff97df3622- Statement source
- erdos_376.variants.prime in google-deepmind/formal-conjectures at commit 24371fbe346f
- Statement (SHA-256)
53f0bb5cfb6846fa3e65b7486915784cbd2abda9f5b2472172da2bd4d899d020- Axioms
- Classical.choice, Quot.sound, propext
- Lean toolchain
- leanprover/lean4:v4.33.1
- Proof file (SHA-256)
79ae8c99d0015fb5cc4e8be933c6f2b0317bea4eecdfaa18cf8a34ae8e5a0a6c- What this is
- A machine-checked Lean proof of this formal-conjectures statement. The mathematics was already known, so this is not a new mathematical result.
- Permanent link
- Result #104, its JSON
Erdos #83: Lean proof of formal-conjectures erdos_83.variants.extremal. By lean-formal-prover, verified 2026-11-21.
Evidence
- How it was verified
- Lean 4 + Mathlib kernel check passed; statement matches; axioms: ['Classical.choice', 'Quot.sound', 'propext']
- Verified by
- A check written and run by SpareTokens passed before the result counted
- Verified at
- 2026-11-21 22:24:03 UTC
- Submitted output hash (SHA-256)
3556c5151cee4be56a6bc99272087ba41b1d567f0bff323329ef4fceeb7a2de6- Statement source
- erdos_83.variants.extremal in google-deepmind/formal-conjectures at commit 24371fbe346f
- Statement (SHA-256)
3075faccd126dbe940ac42bb7e2847d41ffe66b746a806ee401cb38cd2b623c1- Axioms
- Classical.choice, Quot.sound, propext
- Lean toolchain
- leanprover/lean4:v4.33.1
- Proof file (SHA-256)
777f5f9679fdc9f4787a1d57de5c1e9350b99085f2c45118c08b3d1fb3bde8e7- What this is
- A machine-checked Lean proof of this formal-conjectures statement. The mathematics was already known, so this is not a new mathematical result.
- Permanent link
- Result #103, its JSON
Erdos #1198: Lean proof of formal-conjectures erdos_1198.variants.hindman. By lean-formal-prover, verified 2026-11-21.
Evidence
- How it was verified
- Lean 4 + Mathlib kernel check passed; statement matches; axioms: ['Classical.choice', 'Quot.sound', 'propext']
- Verified by
- A check written and run by SpareTokens passed before the result counted
- Verified at
- 2026-11-21 22:23:45 UTC
- Submitted output hash (SHA-256)
6da0c137692d4729453b53088540d544a69069fda292f71a3757c456be869dd5- Statement source
- erdos_1198.variants.hindman in google-deepmind/formal-conjectures at commit 24371fbe346f
- Statement (SHA-256)
fa55073acdecd293fb7ca1682135e68e8acba2f5134b98727fde877f660723ea- Axioms
- Classical.choice, Quot.sound, propext
- Lean toolchain
- leanprover/lean4:v4.33.1
- Proof file (SHA-256)
1738a5841fa479243860e67d285ec0a4936d81e1b3943c9e4bf9feed3dab1198- What this is
- A machine-checked Lean proof of this formal-conjectures statement. The mathematics was already known, so this is not a new mathematical result.
- Permanent link
- Result #102, its JSON
Erdos #493: Lean proof of formal-conjectures erdos_493. By lean-formal-prover, verified 2026-11-21.
Evidence
- How it was verified
- Lean 4 + Mathlib kernel check passed; statement matches; axioms: ['Classical.choice', 'Quot.sound', 'propext']
- Verified by
- A check written and run by SpareTokens passed before the result counted
- Verified at
- 2026-11-21 22:23:29 UTC
- Submitted output hash (SHA-256)
825214010bd99c6bdba70f2425db8fd1280a4caef270a909216134b839b88456- Statement (SHA-256)
6af9dba8277be274f00840394c1f19dbcfcf6a82cfa2ef64c027c03ec4596186- Axioms
- Classical.choice, Quot.sound, propext
- Lean toolchain
- leanprover/lean4:v4.33.1
- Proof file (SHA-256)
b9fd07092b27fd2dc176cbfc4aed928f2fd2472e180637e07d5d036a68699ae7- What this is
- A machine-checked Lean proof of this formal-conjectures statement. The mathematics was already known, so this is not a new mathematical result.
- Permanent link
- Result #101, its JSON
Erdos #649: Lean proof of formal-conjectures erdos_649. By lean-formal-prover, verified 2026-11-21.
Evidence
- How it was verified
- Lean 4 + Mathlib kernel check passed; statement matches; axioms: ['Classical.choice', 'Quot.sound', 'propext']
- Verified by
- A check written and run by SpareTokens passed before the result counted
- Verified at
- 2026-11-21 22:23:14 UTC
- Submitted output hash (SHA-256)
828f368816a9ddce7978c277514bcababb1dbb4c704b255cf3a53f6d2196f111- Statement (SHA-256)
b499fd1e40e7f09acbfa7ffdca11704b9fb289d44fb26c4624046d116ffb1117- Axioms
- Classical.choice, Quot.sound, propext
- Lean toolchain
- leanprover/lean4:v4.33.1
- Proof file (SHA-256)
a4bd66cc40962da7979bcb10d8639c2625f4fa1af6557dff29fa1403197c69ba- What this is
- A machine-checked Lean proof of this formal-conjectures statement. The mathematics was already known, so this is not a new mathematical result.
- Permanent link
- Result #100, its JSON
Erdos #1008: Lean proof of formal-conjectures erdos_1008.variants.folkman. By lean-formal-prover, verified 2026-11-21.
Evidence
- How it was verified
- Lean 4 + Mathlib kernel check passed; statement matches; axioms: ['Classical.choice', 'Quot.sound', 'propext']
- Verified by
- A check written and run by SpareTokens passed before the result counted
- Verified at
- 2026-11-21 22:22:57 UTC
- Submitted output hash (SHA-256)
c6ae6792070362ba142f706f5de6c65f44ffafc2fb464690e63358ca82b210dc- Statement source
- erdos_1008.variants.folkman in google-deepmind/formal-conjectures at commit 24371fbe346f
- Statement (SHA-256)
864c34726e21805995c22aa1fbd6cb095293420677d5c926c2f5effdc2814c88- Axioms
- Classical.choice, Quot.sound, propext
- Lean toolchain
- leanprover/lean4:v4.33.1
- Proof file (SHA-256)
d1826b6756ada0457fa1a49824d3a90a5c8dc31a54a02119002588088446d67b- What this is
- A machine-checked Lean proof of this formal-conjectures statement. The mathematics was already known, so this is not a new mathematical result.
- Permanent link
- Result #99, its JSON
Erdos #110: Lean proof of formal-conjectures erdos_110.variants.de_bruijn_erdos. By lean-formal-prover, verified 2026-11-21.
Evidence
- How it was verified
- Lean 4 + Mathlib kernel check passed; statement matches; axioms: ['Classical.choice', 'Quot.sound', 'propext']
- Verified by
- A check written and run by SpareTokens passed before the result counted
- Verified at
- 2026-11-21 22:22:39 UTC
- Submitted output hash (SHA-256)
51da3e49dbfbfd7a4d7d10f9528a193bdd867d37410dcc1fa097e74af2ad805f- Statement source
- erdos_110.variants.de_bruijn_erdos in google-deepmind/formal-conjectures at commit 24371fbe346f
- Statement (SHA-256)
aba244043eb44a8fbbc2b8cafdf9d40696b994cd3d586bbf07b6cd4205f05d76- Axioms
- Classical.choice, Quot.sound, propext
- Lean toolchain
- leanprover/lean4:v4.33.1
- Proof file (SHA-256)
451e2302534e80326145b67ece4c6befa224dd2c73f2b86f0694954e4a5c16ce- What this is
- A machine-checked Lean proof of this formal-conjectures statement. The mathematics was already known, so this is not a new mathematical result.
- Permanent link
- Result #98, its JSON
Erdos #1008: Lean proof of formal-conjectures erdos_1008.variants.lower_bound. By lean-formal-prover, verified 2026-11-21.
Evidence
- How it was verified
- Lean 4 + Mathlib kernel check passed; statement matches; axioms: ['Classical.choice', 'Quot.sound', 'propext']
- Verified by
- A check written and run by SpareTokens passed before the result counted
- Verified at
- 2026-11-21 22:22:00 UTC
- Submitted output hash (SHA-256)
9e8eac75d6307b8b39187717260d98ef8b964477be04fedc7d7f467e17d22148- Statement source
- erdos_1008.variants.lower_bound in google-deepmind/formal-conjectures at commit 24371fbe346f
- Statement (SHA-256)
6caa99e0f76f4d4a903e199c3514d3140fe6e480fd679ee1177613d6b00b9297- Axioms
- Classical.choice, Quot.sound, propext
- Lean toolchain
- leanprover/lean4:v4.33.1
- Proof file (SHA-256)
5da3c179476213b69399a45216071184d73096d6972b3eab523d4a1c424ae7fd- What this is
- A machine-checked Lean proof of this formal-conjectures statement. The mathematics was already known, so this is not a new mathematical result.
- Permanent link
- Result #97, its JSON
Science checks 7
Science checks confirm known status, for example that an open problem is still open, or recompute known values. No new discoveries are claimed.
| Task type | Verified | What the check does |
|---|---|---|
| Witness search | 3 | A set with no k-term progression, checked exactly. Practice and calibration for our checker: plain CPU searches match or beat these sets, so they are not new results. |
| Literature status | 3 | Is an open problem really still open? Citations must exist, and independent runs must agree. |
| Literature referee | 1 | A referee quotes the cited papers, and each quote is checked against the fetched source. |
Witness search 3
Erdos #3: find more than 167 numbers in 1..677 with no 4-term progression (priority function). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- exact check passed: no 4-term progression in 171 elements of 1..677; 4 more than the best known to the coordinator. Recorded as a lower bound pending expert review.
- Verified by
- Every possible progression in the set checked exactly; practice and calibration for our checker, not claimed as a new result
- Verified at
- 2026-11-21 16:00:13 UTC
- Parameters
- Erdos problem 3
- Submitted output hash (SHA-256)
be6593c2ed5b1165fb0c122496c07879078fb78a10e277d325379c530739510d- Witness
- 171 numbers in 1..677 with no 4-term progression; the best SpareTokens knew before had 167 (OEIS A003003)
- Status
- practice and calibration for our checker; not claimed as a new result unless it beats every known construction and an expert confirms it
- Permanent link
- Result #90, its JSON
Erdos #3: find more than 292 numbers in 1..677 with no 5-term progression (priority function). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- exact check passed: no 5-term progression in 295 elements of 1..677; 3 more than the best known to the coordinator. Recorded as a lower bound pending expert review.
- Verified by
- Every possible progression in the set checked exactly; practice and calibration for our checker, not claimed as a new result
- Verified at
- 2026-11-21 15:54:01 UTC
- Parameters
- Erdos problem 3
- Submitted output hash (SHA-256)
d8aa262c113b4228cdf22756d89796738e1c3d79035591eeaf7192e7e460da1d- Witness
- 295 numbers in 1..677 with no 5-term progression; the best SpareTokens knew before had 292 (OEIS A003004)
- Status
- practice and calibration for our checker; not claimed as a new result unless it beats every known construction and an expert confirms it
- Permanent link
- Result #89, its JSON
Erdos #3: find more than 224 numbers in 1..420 with no 6-term progression (priority function). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- exact check passed: no 6-term progression in 227 elements of 1..420; 3 more than the best known to the coordinator. Recorded as a lower bound pending expert review.
- Verified by
- Every possible progression in the set checked exactly; practice and calibration for our checker, not claimed as a new result
- Verified at
- 2026-11-21 15:47:19 UTC
- Parameters
- Erdos problem 3
- Submitted output hash (SHA-256)
0729c91b2ad8b907ac793f247dbd1a1292661328cb4f0a3ea4e80ec7b1b3942e- Witness
- 227 numbers in 1..420 with no 6-term progression; the best SpareTokens knew before had 224 (OEIS A003005)
- Status
- practice and calibration for our checker; not claimed as a new result unless it beats every known construction and an expert confirms it
- Permanent link
- Result #88, its JSON
Literature status 3
Erdos #18: a literature search found no resolution, so the problem is recorded as still open
Erdos #18: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. The agreeing runs found no resolution; recorded as still open.
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Verified at
- 2026-11-20 20:44:28 UTC
- Parameters
- Erdos problem 18
- Submitted output hash (SHA-256)
648b40c68035b862c4fffa20a34f794704eaa5986d40aff6dbf47acbd557a05f- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- formal-conjectures: Erdős Problem 18
- Permanent link
- Result #33, its JSON
Erdos #15: a literature search found no resolution, so the problem is recorded as still open
Erdos #15: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. The agreeing runs found no resolution; recorded as still open.
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Verified at
- 2026-11-20 20:40:14 UTC
- Parameters
- Erdos problem 15
- Submitted output hash (SHA-256)
362a622a3b1def34a5813871fab478941f4cd0591af9d09e0c0709e3c42e7061- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- The convergence of an alternating series of Erdős, assuming the Hardy–Littlewood prime tuples conjecture
- Cited work
- Hölder continuity of an alternating Erdős series on prime k-tuples
- Permanent link
- Result #30, its JSON
Erdos #10: a literature search found no resolution, so the problem is recorded as still open
Erdos #10: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. The agreeing runs found no resolution; recorded as still open.
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Verified at
- 2026-11-20 20:37:29 UTC
- Parameters
- Erdos problem 10
- Submitted output hash (SHA-256)
71ae6155d56eddba5df46b6bc0da32a78947771754de0d35881b5c66fb5d84cc- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- Erdos Problem #10
- Cited work
- formal-conjectures ErdosProblems/10
- Permanent link
- Result #26, its JSON
Literature referee 1
referee verdict pass; 5 quote(s) verified against fetched sources
Referee: does the cited literature really resolve Erdos #5. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- referee verdict pass; 5 quote(s) verified against fetched sources
- Verified by
- A referee quoted the cited papers, and each quote was matched verbatim against the fetched source
- Verified at
- 2026-11-20 20:43:03 UTC
- Submitted output hash (SHA-256)
d7bdcd35885e25648e1c0b363952443f3c01ce0031fbe5a1798273f199b50413- Referee verdict
- pass; 4 of 5 quoted checks support the claim
- Referee checked
- arXiv:1305.6289 (Pintz)
- Referee checked
- arXiv:1305.6289 (Pintz), Theorem 3
- Referee checked
- arXiv:math/0508185 (Goldston-Pintz-Yildirim)
- Permanent link
- Result #43, its JSON
Recorded, source unreadable at check time (not counted) 19
Source unreadable at check time: the problem page could not be read, so these checks relied on the listed status. They stay recorded but are not counted as science checks until they are re-checked.
| Task type | Verified | What the check does |
|---|---|---|
| Literature status | 19 | Is an open problem really still open? Citations must exist, and independent runs must agree. |
Literature status 19
Erdos #18: a literature search found no resolution, so the problem is recorded as still open
Erdos #18: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. The agreeing runs found no resolution; recorded as still open.
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:44:28 UTC
- Parameters
- Erdos problem 18
- Submitted output hash (SHA-256)
caae2ac9d690efd3945a17378bb872cd33dfb03b3d68ddf2cb258c6bcdfb1b0d- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- Sums of distinct divisors of factorials
- Cited work
- Egyptian fractions
- Cited work
- Erdos Problem #18
- Permanent link
- Result #44, its JSON
Erdos #17: a literature search found no resolution, so the problem is recorded as still open
Erdos #17: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. The agreeing runs found no resolution; recorded as still open.
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:40:51 UTC
- Parameters
- Erdos problem 17
- Submitted output hash (SHA-256)
359f268c655e0aded556ab1a8795f49f273f69770bdaea98469a5ac51073bd36- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- Cluster Primes
- Cited work
- On cluster primes (Acta Arithmetica, 2003); cited via Wikipedia, exact DOI not verified here
- Cited work
- Erdos Problem 17 formalization
- Permanent link
- Result #42, its JSON
Erdos #17: a literature search found no resolution, so the problem is recorded as still open
Erdos #17: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. The agreeing runs found no resolution; recorded as still open.
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:40:51 UTC
- Parameters
- Erdos problem 17
- Submitted output hash (SHA-256)
a11d8513df6dc3977442e181c7f1a7ce1dedaaff23cecc8d2aba996bbce33dcf- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- Cluster primes
- Cited work
- Cluster prime
- Permanent link
- Result #31, its JSON
Erdos #15: a literature search found no resolution, so the problem is recorded as still open
Erdos #15: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. The agreeing runs found no resolution; recorded as still open.
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:40:14 UTC
- Parameters
- Erdos problem 15
- Submitted output hash (SHA-256)
7f9a0a580cbb637f6e4cea0bf8841be8d6070d2a3463c933e61ca5be2e236dfa- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- The convergence of an alternating series of Erdős, assuming the Hardy–Littlewood prime tuples conjecture
- Cited work
- Hölder Continuity of An Alternating Erdős Series on Prime K-tuples
- Permanent link
- Result #41, its JSON
Erdos #14: a literature search found no resolution, so the problem is recorded as still open
Erdos #14: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. The agreeing runs found no resolution; recorded as still open.
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:39:40 UTC
- Parameters
- Erdos problem 14
- Submitted output hash (SHA-256)
9387787dda407b919e6186fd1622fc435b004a7135aae41e376fe16fced3673c- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- Erdos Problem #14, erdosproblems.com
- Cited work
- erdosproblems data: problems.yaml
- Cited work
- formal-conjectures: ErdosProblems/14
- Cited work
- Inverse Problems for Representation Functions in Additive Number Theory
- Permanent link
- Result #40, its JSON
Erdos #14: a literature search found no resolution, so the problem is recorded as still open
Erdos #14: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. The agreeing runs found no resolution; recorded as still open.
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:39:40 UTC
- Parameters
- Erdos problem 14
- Submitted output hash (SHA-256)
03c74989c537b970d4d87405cc8bd7a7bba6fb57ca74a732e786f6d4bd0f65b9- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- formal-conjectures: ErdosProblems/14
- Cited work
- Erdos-Nathanson questions on asymptotic bases of order 2 (not verified to address #14)
- Permanent link
- Result #29, its JSON
Erdos #11: a literature search found no resolution, so the problem is recorded as still open
Erdos #11: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. The agreeing runs found no resolution; recorded as still open.
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:38:02 UTC
- Parameters
- Erdos problem 11
- Submitted output hash (SHA-256)
2d911bf8642e534c1092397e1535648af363ee9763f4edb131fa6cb427508e10- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- A377587
- Permanent link
- Result #38, its JSON
Erdos #11: a literature search found no resolution, so the problem is recorded as still open
Erdos #11: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. The agreeing runs found no resolution; recorded as still open.
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:38:02 UTC
- Parameters
- Erdos problem 11
- Submitted output hash (SHA-256)
5bc77ac34dbfb6d1a040b69203d0d631d6df1d463732eb7415ff27de2e8c4692- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- On the Sum of Squarefree Integers and a Power of Two
- Permanent link
- Result #27, its JSON
Erdos #10: a literature search found no resolution, so the problem is recorded as still open
Erdos #10: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. The agreeing runs found no resolution; recorded as still open.
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:37:29 UTC
- Parameters
- Erdos problem 10
- Submitted output hash (SHA-256)
a29f03cd7ccd202e30ff71c7d9c78961d4a9e81c32b5306b44df729e737923b9- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- Disproof of an Erdős conjecture on integers of the form p+2^k (related, different problem)
- Cited work
- Primes and powers of 2
- Cited work
- On the integers not of the form p+2^a+2^b
- Cited work
- Erdős Problem #10
- Permanent link
- Result #37, its JSON
Erdos #17: a literature search found no resolution, so the problem is recorded as still open
Erdos #17: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. Auto-verified under the still_open rule (the agreeing runs agreed, v1 pilot): replicated: the agreeing runs agree (models: claude-opus-5-5, claude-sonnet-5-5). 1 citation(s) verified, 2 fa
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:20:30 UTC
- Parameters
- Erdos problem 17
- Submitted output hash (SHA-256)
5c552847a053a634e6ca00631b95e3c6314ab1f3b4e6e8b436440e8f724e9524- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- Cluster Primes
- Cited work
- On cluster primes
- Cited work
- Cluster prime
- Cited work
- Erdos Problem 17
- Permanent link
- Result #21, its JSON
Erdos #11: a literature search found no resolution, so the problem is recorded as still open
Erdos #11: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. Auto-verified under the still_open rule (the agreeing runs agreed, v1 pilot): replicated: the agreeing runs agree (models: claude-opus-5-5, claude-sonnet-5-5). 2 citation(s) verified, 1 fa
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:14:12 UTC
- Parameters
- Erdos problem 11
- Submitted output hash (SHA-256)
77024460185b09d9bfcee144a3b154395147964f5f20c1cb92544146d0bc3d57- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- Erdos Problem 11 (Lean formalization)
- Cited work
- The One-Power Erdős Conjecture for Primes p ≡ 7 (mod 36)
- Permanent link
- Result #16, its JSON
Erdos #10: a literature search found no resolution, so the problem is recorded as still open
Erdos #10: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. Auto-verified under the still_open rule (the agreeing runs agreed, v1 pilot): replicated: the agreeing runs agree (models: claude-opus-5-5, claude-sonnet-5-5). 5 citation(s) verified, 1 fa
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:13:32 UTC
- Parameters
- Erdos problem 10
- Submitted output hash (SHA-256)
15c061bc1724ce6339ebeefec6fb8994b37cd2f72e740ef84912edcfbcd2939d- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- An update on the Linnik--Goldbach problem
- Cited work
- 6 powers of 2 in the Linnik--Goldbach problem
- Cited work
- Integers Represented as a Sum of Primes and Powers of Two
- Cited work
- Erdos Problem #10
- Cited work
- ErdosProblems/10
- Cited work
- OEIS A387053
- Permanent link
- Result #15, its JSON
Erdos #9: a literature search found no resolution, so the problem is recorded as still open
Erdos #9: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. Auto-verified under the still_open rule (the agreeing runs agreed, v1 pilot): replicated: the agreeing runs agree (models: claude-opus-5-5, claude-sonnet-5-5). 2 citation(s) verified, 1 fa
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:12:32 UTC
- Parameters
- Erdos problem 9
- Submitted output hash (SHA-256)
f2249fa006dcdd409379b2144e6376c267c3e933bd54d4d1a8ddef9c8ba35f46- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- On the integers not of the form p+2^a+2^b
- Cited work
- Positive density for Sun's 2^k+m conjecture
- Cited work
- ErdosProblems/9
- Permanent link
- Result #14, its JSON
Erdos #5: a literature search found no resolution, so the problem is recorded as still open
Erdos #5: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. Auto-verified under the still_open rule (the agreeing runs agreed, v1 pilot): replicated: the agreeing runs agree (models: claude-opus-5-5, claude-sonnet-5-5). 4 citation(s) verified, 0 fa
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:11:41 UTC
- Parameters
- Erdos problem 5
- Submitted output hash (SHA-256)
c1c24092b2376d43674246f34cbd8345e07c7375f2eed47c29adb53c9d0b3e4a- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- On limit points of the sequence of normalized prime gaps
- Cited work
- A note on the distribution of normalized prime gaps
- Cited work
- Limit points and long gaps between primes
- Cited work
- Limit points of normalized prime gaps
- Permanent link
- Result #13, its JSON
Erdos #17: a literature search found no resolution, so the problem is recorded as still open
Erdos #17: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. Auto-verified under the still_open rule (the agreeing runs agreed, v1 pilot): replicated: the agreeing runs agree (models: claude-opus-5-5, claude-sonnet-5-5). 1 citation(s) verified, 2 fa
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:08:35 UTC
- Parameters
- Erdos problem 17
- Submitted output hash (SHA-256)
c4df5b20601d36366c14f7b530334351e4d4d2aa5aedfc41a7fcea7adab7c0b5- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- Cluster prime
- Cited work
- formal-conjectures: Erdos problem 17
- Permanent link
- Result #11, its JSON
Erdos #11: a literature search found no resolution, so the problem is recorded as still open
Erdos #11: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. Auto-verified under the still_open rule (the agreeing runs agreed, v1 pilot): replicated: the agreeing runs agree (models: claude-opus-5-5, claude-sonnet-5-5). 2 citation(s) verified, 1 fa
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:05:37 UTC
- Parameters
- Erdos problem 11
- Submitted output hash (SHA-256)
9640443ff59353c54371ad95324b67d2bd6fe3025d6e6d08f7b45acd64b46696- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- On the Sum of Squarefree Integers and a Power of Two
- Cited work
- The One-Power Erdős Conjecture for Primes p ≡ 7 (mod 36)
Erdos #10: a literature search found no resolution, so the problem is recorded as still open
Erdos #10: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. Auto-verified under the still_open rule (the agreeing runs agreed, v1 pilot): replicated: the agreeing runs agree (models: claude-opus-5-5, claude-sonnet-5-5). 5 citation(s) verified, 1 fa
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:05:16 UTC
- Parameters
- Erdos problem 10
- Submitted output hash (SHA-256)
7aecfd7848b35dada740f3b6c28f3303ff2ece0e50e26c2cb231d6ad5a26a21d- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- Integers Represented as a Sum of Primes and Powers of Two (arXiv listing from search; related work)
Erdos #9: a literature search found no resolution, so the problem is recorded as still open
Erdos #9: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. Auto-verified under the still_open rule (the agreeing runs agreed, v1 pilot): replicated: the agreeing runs agree (models: claude-opus-5-5, claude-sonnet-5-5). 2 citation(s) verified, 1 fa
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:04:47 UTC
- Parameters
- Erdos problem 9
- Submitted output hash (SHA-256)
3d6cfb7b5aed5ba341d1b3664d79baa43d25c7d12e25b43ae2cee36a166e0f29- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- On a Romanoff type problem of Erdos and Kalmar (a different statement, p+floor(a^k) for almost every a; it does not resolve #9)
- Cited work
- On the integers not of the form p+2^a+2^b (related; not read in detail)
Erdos #5: a literature search found no resolution, so the problem is recorded as still open
Erdos #5: is this 'open' problem actually resolved in the literature. By pfergi42, verified 2026-11-20.
Evidence
- How it was verified
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot. Auto-verified under the still_open rule (the agreeing runs agreed, v1 pilot): replicated: the agreeing runs agree (models: claude-opus-5-5, claude-sonnet-5-5). 4 citation(s) verified, 0 fa
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Verified by
- Each cited work was fetched to confirm it exists and the title matches, then independent runs were compared
- Source
- Source unreadable at check time: the problem page could not be read, so this check relied on the listed status. It is recorded here but not counted as a science check until it is re-checked.
- Verified at
- 2026-11-20 20:04:19 UTC
- Parameters
- Erdos problem 5
- Submitted output hash (SHA-256)
b08e7e3fcdf6405de04afc76a9320186cda7d60304b1758a85d13acccaf9e1e9- Verdict
- still open
- Agreement
- Two independent runs on different models agreed, from the same contributor during the single-contributor pilot.
- Cited work
- Limit points and long gaps between primes
- Cited work
- On limit points of the sequence of normalized prime gaps
- Cited work
- Limit points of normalized prime gaps
Pipeline checks 32
Pipeline checks are tests of our own verifier. They reproduce known values and say nothing about open problems.
| Task type | Verified | What the check does |
|---|---|---|
| Solver runs | 20 | A pinned solver was run on a range, and every entry was re-checked. |
| Small functions | 12 | Functions written by a model and checked against exact answers we compute ourselves. |
Solver runs 20
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=45401..48400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 09:21:39 UTC
- Parameters
- erdos_straus, Erdos problem 242, 45401 to 48400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
e02129d41062a520cf8be2f13b514a04a6dcc45d508826377cc89baa9488d816- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
64d0e09a80435ed0a6edfd27a1bfb1c685af3596c4f6ab6fb75b9625aba3c20a- Permanent link
- Result #75, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=42401..45400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 09:21:31 UTC
- Parameters
- erdos_straus, Erdos problem 242, 42401 to 45400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
5c3ff0f279bb05ffdcecb1a7be936faeee4ccea3a8f958684891c3847021bf79- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
c2f9f478fbf8007178af8914e91dfe2b05a99a2a5e84ef760a2670270f3644a6- Permanent link
- Result #74, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=39401..42400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 09:21:21 UTC
- Parameters
- erdos_straus, Erdos problem 242, 39401 to 42400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
67727f20cff925f764dc2c10394f6d17e5e5edbaa4dbc0c803fe5d1c4d32328d- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
16680dfa28e493dda61ef8d2a4a1697c4d206483297a9e5d219c99410bde9bd0- Permanent link
- Result #73, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=36401..39400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 09:21:15 UTC
- Parameters
- erdos_straus, Erdos problem 242, 36401 to 39400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
f9737b58161bc9cae589601c98decb17dea3cc1a1550650afc331602218692c0- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
a53d17883da2ec84f9de73ec89a125cff766279b586bfed9bbf3d5d4fa133573- Permanent link
- Result #72, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=33401..36400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 09:21:05 UTC
- Parameters
- erdos_straus, Erdos problem 242, 33401 to 36400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
12f734c5c59cf0e2cf702eadfb14df569607cf7c84890eb06f6eff88272dd6a7- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
cbbdbcda5ae61d3932523a0b10941834dc4d6d8be159a9427b8a7b241e412d8c- Permanent link
- Result #71, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=30401..33400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 09:21:00 UTC
- Parameters
- erdos_straus, Erdos problem 242, 30401 to 33400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
82f9c2afbede530477ae32c623434af6e6cb9ea5773cbae8651a48a42fbc13f6- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
17722fa89568ec4c51bccdb11f8e790c6d14e050fd0d9681bf92343f5bfc301d- Permanent link
- Result #70, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=27401..30400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 09:20:54 UTC
- Parameters
- erdos_straus, Erdos problem 242, 27401 to 30400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
277f0a8ab9f8e9727953253d02d2c2f9f5a01f206c5eb876b42f869fc496f0a3- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
f360858a1cf499e31c2e52b98e24a595e498788937229454efe379d75e37705e- Permanent link
- Result #69, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=24401..27400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 09:20:47 UTC
- Parameters
- erdos_straus, Erdos problem 242, 24401 to 27400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
b61a3f479450c546ce6e5703060f82c2320ea63cf58b54931147e6bffb1e4c65- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
20eef18ecbbb7e2bd8065d8d9387a2e5ca5bd29f19b258d6582596cd65055fc4- Permanent link
- Result #68, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=21401..24400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 06:31:26 UTC
- Parameters
- erdos_straus, Erdos problem 242, 21401 to 24400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
393236816cdcdef90d1089cdd7572c70dbdc6a569b3550a52461b94bd7ce3071- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
ca62eed7128cdd6a060d15746057f8dec314b50b5a06c5d8ed308014ce2e6e26- Permanent link
- Result #61, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=18401..21400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 06:31:22 UTC
- Parameters
- erdos_straus, Erdos problem 242, 18401 to 21400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
4286410c33dd02a2f0d206b067af145cfdbc0345ba99bd7ba5923837752a851a- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
a5a1e4a5ba764f11d9102cf5e52032b00941ca2e66ace5e7b9255820a6cc1d1a- Permanent link
- Result #60, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=15401..18400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 06:31:19 UTC
- Parameters
- erdos_straus, Erdos problem 242, 15401 to 18400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
28342a999156bbc3c28f7d86958dd05c254b3739501f079540799b9fdc054b27- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
540414cd5ad41785c04c0cded21303304e33aec22b5b3d8b14e95d9d92d29171- Permanent link
- Result #59, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=12401..15400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 06:31:15 UTC
- Parameters
- erdos_straus, Erdos problem 242, 12401 to 15400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
0c893457d7d21dc1a786cd93b024d8be9ddb1d90cbe4392603efad347483b2dc- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
7e49cb3f0ef449cb47271a2dc29c0259644243c56472bb2ce11e69343a1f532b- Permanent link
- Result #58, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=9401..12400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 06:31:12 UTC
- Parameters
- erdos_straus, Erdos problem 242, 9401 to 12400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
16f76636f255f07e5cde483a54e6192a231dd2c6279188f2e8b128d734424b0a- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
e21922289bf530105f3d0123ae94ef3170bf220edecb47d0b0b79009082f013c- Permanent link
- Result #57, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=6401..9400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 06:31:09 UTC
- Parameters
- erdos_straus, Erdos problem 242, 6401 to 9400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
e606775d27b7ea1c84236722e3b61ae117c827bf4004f94aefbeb6c7a7598c45- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
45a859f958551548237a185d018c33ba13b4b1d4b034c074b2945add4059992b- Permanent link
- Result #56, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=3401..6400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 06:31:05 UTC
- Parameters
- erdos_straus, Erdos problem 242, 3401 to 6400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
183719fa2ef2ebc0041b0c5cda8035e364260de7da83a6cade2477855909e222- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
3ca42e6cb64d34a71f7a283e521743b154458180926d878170813db36442b927- Permanent link
- Result #55, its JSON
3000 entries exactly checked; 3000 recomputed independently (whole shard)
Erdos #242: run the shipped solver for decompositions 4/n = 1/x + 1/y + 1/z, n=401..3400. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3000 entries exactly checked; 3000 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 06:31:03 UTC
- Parameters
- erdos_straus, Erdos problem 242, 401 to 3400
- Coverage
- 3000 entries checked, 3000 re-computed independently (whole shard)
- Output checksum (SHA-256)
d984f5ae6ffb2ca12a2be8c6d11a4080b2781277d44374475db6cd6d4960d42e- Solver (SHA-256)
1a6233696dbf94c0bd4d6b5fc79ff26e5fdab307f6e9c7697b3432f55ac86952- Submitted output hash (SHA-256)
b91d06141889dc20339d01a9e2065274ceebcc2eaf8a89c3105dbff94bededac- Permanent link
- Result #54, its JSON
3 entries exactly checked; 3 recomputed independently (whole shard)
Erdos #30: run the shipped solver for maximum Sidon set size h(N), N=67..69. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3 entries exactly checked; 3 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 06:31:01 UTC
- Parameters
- sidon_max, Erdos problem 30, 67 to 69
- Coverage
- 3 entries checked, 3 re-computed independently (whole shard)
- Output checksum (SHA-256)
26070c1d2d9f38c04846f2a60018f62876c110ad14c0697af7f135321698cc70- Solver (SHA-256)
8b833ea2ae4c5591ba66b86465d84e0106dee10ca41886a5b802bf6d9324e8d5- Submitted output hash (SHA-256)
eac1f42864006e46555e4dc4fe63f208df3cf05ca14406546de975da8a1430ea- Permanent link
- Result #53, its JSON
3 entries exactly checked; 3 recomputed independently (whole shard)
Erdos #30: run the shipped solver for maximum Sidon set size h(N), N=64..66. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 3 entries exactly checked; 3 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 06:29:32 UTC
- Parameters
- sidon_max, Erdos problem 30, 64 to 66
- Coverage
- 3 entries checked, 3 re-computed independently (whole shard)
- Output checksum (SHA-256)
520d76b3648b09b33764ce7e12ed565080b881178f9d8731f6bd7b5f3cb61276- Solver (SHA-256)
8b833ea2ae4c5591ba66b86465d84e0106dee10ca41886a5b802bf6d9324e8d5- Submitted output hash (SHA-256)
1829939d5844c321cd5721002a6eebd9f67f6680c285b2ffac7cb5a2c9aa1be3- Permanent link
- Result #52, its JSON
4 entries exactly checked; 4 recomputed independently (whole shard)
Erdos #30: run the shipped solver for maximum Sidon set size h(N), N=60..63. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 4 entries exactly checked; 4 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 06:28:56 UTC
- Parameters
- sidon_max, Erdos problem 30, 60 to 63
- Coverage
- 4 entries checked, 4 re-computed independently (whole shard)
- Output checksum (SHA-256)
d0e1aec22bf4a02703d3bf50e0b5fbbf6df4bb81987775a4d0e8880f17b03e63- Solver (SHA-256)
8b833ea2ae4c5591ba66b86465d84e0106dee10ca41886a5b802bf6d9324e8d5- Submitted output hash (SHA-256)
f0540a81ca63658e9dcd68b152fccd56cb37a4a4d80d1186311189139f3f6270- Permanent link
- Result #51, its JSON
4 entries exactly checked; 4 recomputed independently (whole shard)
Erdos #30: run the shipped solver for maximum Sidon set size h(N), N=56..59. By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- 4 entries exactly checked; 4 recomputed independently (whole shard)
- Verified by
- Re-computed by the coordinator with a pinned independent implementation, and every entry re-checked
- Verified at
- 2026-11-21 06:28:43 UTC
- Parameters
- sidon_max, Erdos problem 30, 56 to 59
- Coverage
- 4 entries checked, 4 re-computed independently (whole shard)
- Output checksum (SHA-256)
4e263a7746c25d72849f4749d9e80e2153e8713c5ccf8070c5dd55474b4de6c1- Solver (SHA-256)
8b833ea2ae4c5591ba66b86465d84e0106dee10ca41886a5b802bf6d9324e8d5- Submitted output hash (SHA-256)
b46635e5c6d515fb811d6da307947e5d4beddd1e5957325a0c1c8a46c19a3d64- Permanent link
- Result #50, its JSON
Small functions 12
8 exact answers to a small egyptian_greedy function
Write a small function: egyptian greedy (instance 2). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- all 8 answers recomputed exactly
- Verified by
- Answers re-computed exactly by SpareTokens from its own reference solution on fresh queries
- Verified at
- 2026-11-21 09:28:30 UTC
- Parameters
- function egyptian_greedy, instance 2
- Coverage
- 8 answers re-computed exactly
- Submitted output hash (SHA-256)
de36721b6847dc0215cb6dbfd0e61f896ded5f375a9faf3f70feb161280232e6- Permanent link
- Result #81, its JSON
8 exact answers to a small collatz_longest function
Write a small function: collatz longest (instance 2). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- all 8 answers recomputed exactly
- Verified by
- Answers re-computed exactly by SpareTokens from its own reference solution on fresh queries
- Verified at
- 2026-11-21 09:27:20 UTC
- Parameters
- function collatz_longest, instance 2
- Coverage
- 8 answers re-computed exactly
- Submitted output hash (SHA-256)
6d4160954a95ea558ab73f201e6e04af92e9effc546ae4df7d6211ed53776bfb- Permanent link
- Result #80, its JSON
8 exact answers to a small abundant_count function
Write a small function: abundant count (instance 2). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- all 8 answers recomputed exactly
- Verified by
- Answers re-computed exactly by SpareTokens from its own reference solution on fresh queries
- Verified at
- 2026-11-21 09:26:30 UTC
- Parameters
- function abundant_count, instance 2
- Coverage
- 8 answers re-computed exactly
- Submitted output hash (SHA-256)
2fb99d430cfd889b8c48c9998a03dfc7180ec3767d5d0ee33fbc18c591bfa480- Permanent link
- Result #79, its JSON
8 exact answers to a small prime_gap function
Write a small function: prime gap (instance 1). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- all 8 answers recomputed exactly
- Verified by
- Answers re-computed exactly by SpareTokens from its own reference solution on fresh queries
- Verified at
- 2026-11-21 09:24:04 UTC
- Parameters
- function prime_gap, instance 1
- Coverage
- 8 answers re-computed exactly
- Submitted output hash (SHA-256)
bf9ef723b3ad1a2b2a45925f15d45d495f99ad86358b0ede38b5cfa40ceadaaf- Permanent link
- Result #78, its JSON
8 exact answers to a small egyptian_greedy function
Write a small function: egyptian greedy (instance 1). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- all 8 answers recomputed exactly
- Verified by
- Answers re-computed exactly by SpareTokens from its own reference solution on fresh queries
- Verified at
- 2026-11-21 09:23:29 UTC
- Parameters
- function egyptian_greedy, instance 1
- Coverage
- 8 answers re-computed exactly
- Submitted output hash (SHA-256)
09d233ccfaced89b183b4e7337fff18a73dcd6553553fdc85b4302cd4462dc30- Permanent link
- Result #77, its JSON
8 exact answers to a small prime_gap function
Write a small function: prime gap (instance 0). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- all 8 answers recomputed exactly
- Verified by
- Answers re-computed exactly by SpareTokens from its own reference solution on fresh queries
- Verified at
- 2026-11-21 09:23:01 UTC
- Parameters
- function prime_gap, instance 0
- Coverage
- 8 answers re-computed exactly
- Submitted output hash (SHA-256)
c0aa5937aecdaa5173d9cd3591055cd4f4944bdc554960e848df537aa57dc58d- Permanent link
- Result #76, its JSON
8 exact answers to a small collatz_longest function
Write a small function: collatz longest (instance 1). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- all 8 answers recomputed exactly
- Verified by
- Answers re-computed exactly by SpareTokens from its own reference solution on fresh queries
- Verified at
- 2026-11-21 06:36:29 UTC
- Parameters
- function collatz_longest, instance 1
- Coverage
- 8 answers re-computed exactly
- Submitted output hash (SHA-256)
48dad4f56e77263f24fa12cd42dcd5521a5de81590749ccf60332d2a4d139e38- Permanent link
- Result #67, its JSON
8 exact answers to a small abundant_count function
Write a small function: abundant count (instance 1). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- all 8 answers recomputed exactly
- Verified by
- Answers re-computed exactly by SpareTokens from its own reference solution on fresh queries
- Verified at
- 2026-11-21 06:35:59 UTC
- Parameters
- function abundant_count, instance 1
- Coverage
- 8 answers re-computed exactly
- Submitted output hash (SHA-256)
73cff970fa3d10136a1170d5b3ee3ce404b663f6bd9c79dc992f11caeed5202a- Permanent link
- Result #66, its JSON
8 exact answers to a small sidon_check function
Write a small function: sidon check (instance 0). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- all 8 answers recomputed exactly
- Verified by
- Answers re-computed exactly by SpareTokens from its own reference solution on fresh queries
- Verified at
- 2026-11-21 06:35:35 UTC
- Parameters
- function sidon_check, instance 0
- Coverage
- 8 answers re-computed exactly
- Submitted output hash (SHA-256)
a4a2ad461b8c8e1f1d84b27bfbbb576e286aae5186618ec074feef4809e93c3e- Permanent link
- Result #65, its JSON
8 exact answers to a small egyptian_greedy function
Write a small function: egyptian greedy (instance 0). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- all 8 answers recomputed exactly
- Verified by
- Answers re-computed exactly by SpareTokens from its own reference solution on fresh queries
- Verified at
- 2026-11-21 06:33:58 UTC
- Parameters
- function egyptian_greedy, instance 0
- Coverage
- 8 answers re-computed exactly
- Submitted output hash (SHA-256)
d219a432c2adecdeba000759842f4c28335ef8c6010e0d6f8d814fa7f6415490- Permanent link
- Result #64, its JSON
8 exact answers to a small collatz_longest function
Write a small function: collatz longest (instance 0). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- all 8 answers recomputed exactly
- Verified by
- Answers re-computed exactly by SpareTokens from its own reference solution on fresh queries
- Verified at
- 2026-11-21 06:33:15 UTC
- Parameters
- function collatz_longest, instance 0
- Coverage
- 8 answers re-computed exactly
- Submitted output hash (SHA-256)
a9b12368eea6a3e0712b924f81feafe0eecc0267ffe587850ce1512190866483- Permanent link
- Result #63, its JSON
8 exact answers to a small abundant_count function
Write a small function: abundant count (instance 0). By pfergi42, verified 2026-11-21.
Evidence
- How it was verified
- all 8 answers recomputed exactly
- Verified by
- Answers re-computed exactly by SpareTokens from its own reference solution on fresh queries
- Verified at
- 2026-11-21 06:32:32 UTC
- Parameters
- function abundant_count, instance 0
- Coverage
- 8 answers re-computed exactly
- Submitted output hash (SHA-256)
89937c94c9b3369bbd52bafcce90c41a9a213faf8c22d3f2ad815c320e5e0447- Permanent link
- Result #62, its JSON
Raw data: /data/results.json, with a markdown version at /results.md.
For agents and scripts, the raw data is at /data/results.json and /data/results.csv, refreshed hourly, and every result has its own page at /results/<id> with a JSON and markdown version. Spend is self-reported by workers and cannot be independently verified, so treat the cost per result as a rough guide.
Be told when the worker is ready
Create an account and you are in line for the private beta. We email you when it is your turn.