Claim

Persistent rumours around the time of Q* and Strawberry held that OpenAI's early reasoning models were trained on, or tested against, objectives of inverting cryptographically secure hash functions such as the SHA suite.

Type OBSERVATIONAL · truth-evaluable: yes · status at snapshot: ACTIVE · created seq 2329

33 words; atomic claims are usually under 25a conjunction joins clauses that may be separately evaluablea comma-separated series may list several propositions

Referenced 278 times: checked 6, looked up by assistants 0, viewed 148, shared 124. A count of attention, not evidence.

Topics

AI · 03192ed8 Security · 03192ed8

Assessment

Unresolved

Score: 0.4641 under ledger-default@0.3.0 at snapshot 27745

3 of 4 review checks done · 3 counted

4 independent evidence lineages · review checks 3 of 4 · stability HIGH

Main unresolved issue: 10 counted evidence items have no independence review; repeats of one origin may be counted as several coverage:abd4fca4-5764-865f-abaf-227324079aa6

Not yet independently audited.Evidence points both ways.A quoted passage from blacklynxbrief.com and stacker.news was not found on the page when Galedra fetched it.A quoted passage is in something this server does not read, such as a PDF or a recording, so it has not been checked here.

Why?

Strongest counted support: A 2023 report documents a Q-Star leak and associated claims that it could break encryption, showing that encryption-breaking rumors circulated around Q*. weight 0.540000 evidence

Strongest counted contradiction: Reuters reports math-benchmark testing and says it could not determine whether that system was Strawberry; it supplies no evidence of SHA-hash inversion. weight 0.459000 evidence

Suppressed as dependent: Speaker reports persistent rumours that early reasoning models were trained on or tested against inverting secure hash functions.; The excerpt reports persistent rumors that early reasoning models around Q-star and Strawberry were trained or tested on inverting cryptographically secure hashes.

Review gaps: independence reviewed

What would most change this: One independent, direct measurement that supports this claim would move it from Unresolved to Supported.

Hide calculation

0.4641 under ledger-default@0.3.0 at snapshot 27745 · state Unresolved · Review checks: 3 of 4 · stability HIGH · 10 sources, 4 independent lineages

Model ledger-default@0.3.0 (the default model, not the answer) · snapshot 27745 · trace sha256:219efecbc08227b0da00a7fcbbd83e1d5582f0617b26d9691a4f625df1750786

Last changed at entry seq 16982; nothing logged since has borne on this claim.

How the number was reached

Each counted link is weighed by multiplying the factors the model declares; a factor it does not declare is not shown. Every factor name below links to what that number means.

Link	Working	Weight	Counted as
fd2b5e3e contradict	relevance MODERATE 0.6 × observation DIRECT_TEXT 0.9 × interpretation 2 steps, 1 − 0.15 × 2 0.700000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance INDEPENDENT 1.0 × edition NAMED_OR_UNRELATED 1	0.378000	0.378000
38abdbb6 qualify	relevance MODERATE 0.6 × observation DIRECT_TEXT 0.9 × interpretation 0 steps, 1 − 0.15 × 0 1.000000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance INDEPENDENT 1.0 × edition NAMED_OR_UNRELATED 1	0.540000	0.000000 (non directional)
f397bae5 support	relevance DIRECT 2.0 × observation HEARSAY 0.1 × interpretation 1 step, 1 − 0.15 × 1 0.850000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance SELF 0.0 × edition NAMED_OR_UNRELATED 1	0.000000	0.000000 (dependent strongest only, kept 9cf6fc6d)
8226a1ea support	relevance MODERATE 0.6 × observation DIRECT_TEXT 0.9 × interpretation 0 steps, 1 − 0.15 × 0 1.000000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance INDEPENDENT 1.0 × edition NAMED_OR_UNRELATED 1	0.540000	0.540000
611fe8d6 qualify	relevance MODERATE 0.6 × observation DIRECT_TEXT 0.9 × interpretation 1 step, 1 − 0.15 × 1 0.850000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance INDEPENDENT 1.0 × edition NAMED_OR_UNRELATED 1	0.459000	0.000000 (non directional)
9cf6fc6d support	relevance WEAK 0.2 × observation HEARSAY 0.1 × interpretation 2 steps, 1 − 0.15 × 2 0.700000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance SELF 0.0 × edition NAMED_OR_UNRELATED 1	0.000000	0.000000
de23d5c7 support	relevance WEAK 0.2 × observation DIRECT_TEXT 0.9 × interpretation 1 step, 1 − 0.15 × 1 0.850000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance INDEPENDENT 1.0 × edition NAMED_OR_UNRELATED 1	0.153000	0.153000
e6c92da6 qualify	relevance MODERATE 0.6 × observation DIRECT_TEXT 0.9 × interpretation 1 step, 1 − 0.15 × 1 0.850000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance INDEPENDENT 1.0 × edition NAMED_OR_UNRELATED 1	0.459000	0.000000 (non directional)
80ae660f contradict	relevance MODERATE 0.6 × observation DIRECT_TEXT 0.9 × interpretation 1 step, 1 − 0.15 × 1 0.850000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance INDEPENDENT 1.0 × edition NAMED_OR_UNRELATED 1	0.459000	0.459000
032ca64f support	relevance DIRECT 2.0 × observation DIRECT_TEXT 0.9 × interpretation 1 step, 1 − 0.15 × 1 0.850000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance SELF 0.0 × edition NAMED_OR_UNRELATED 1	0.000000	0.000000 (dependent strongest only, kept 9cf6fc6d)

prior 0.50 → log-odds 0.000000 + -0.144000 (the counted weights) = -0.144000 → probability 0.4641

Under neighbouring settings of the same model: 0.4748 to 0.4533, a spread of 0.0215.

Half-even rounding, six places for weights and four for the probability. The same inputs under the same model reproduce this exactly; that is what the trace hash above is for. How a score is worked out explains every factor, why it is there, and what the model sets it to.

Your view

Registered views

Nobody has registered a view yet.

Registered views are personal belief, not evidence: they never change Galedra's assessment. Affiliations are self-declared and shown only for groups of 5 or more.

Checked as part of

Statements this claim was broken out of. The claim is scored on its own; a statement's reading only counts its claims.

“Persistent rumours around the time of Q* and Strawberry held that OpenAI's early reasoning models were trained on, or tested against, objectives of inverting...” · 1 claim · 29 September 2026
“Persistent rumours around the time of Q* and Strawberry held that OpenAI's early reasoning models were trained on, or tested against, objectives of inverting...” · 1 claim · 28 September 2026
“Persistent rumours around the time of Q* and Strawberry held that OpenAI's early reasoning models were trained on, or tested against, objectives of inverting...” · 1 claim · 28 September 2026
“Persistent rumours around the time of Q* and Strawberry held that OpenAI's early reasoning models were trained on, or tested against, objectives of inverting...” · 1 claim · 28 September 2026
“Persistent rumours around the time of Q* and Strawberry held that OpenAI's early reasoning models were trained on, or tested against, objectives of inverting...” · 1 claim · 20 September 2026
“The Navier–Stokes proof was produced by roughly 10,000 OpenAI agents working for about 88 hours.” · 32 claims · 20 September 2026

Evidence for

Wissner-Gross reports persistent rumours of early reasoning models aimed at inverting hashes.

From OpenAI Agents Hijack a German Website, Jensen Declares AGI Arrived, and OpenAI Solves Navier-Stokes Read 2026-09-20 Relevance WEAK Steps 2 Counted

Explicitly reported as rumour, with the speaker asking OpenAI for ground truth.
Speaker reports persistent rumours that early reasoning models were trained on or tested against inverting secure hash functions.

From OpenAI Agents Hijack a German Website, Jensen Declares AGI Arrived, and OpenAI Solves Navier-Stokes Read 2026-09-20 Relevance DIRECT Steps 1 Counted

The claim is scoped to what the rumours held, and the passage reports exactly that, so the two match at the same level; the models are tied to OpenAI only via the Q-star and Strawberry framing.
The excerpt reports persistent rumors that early reasoning models around Q-star and Strawberry were trained or tested on inverting cryptographically secure hashes.

From OpenAI Agents Hijack a German Website, Jensen Declares AGI Arrived, and OpenAI Solves Navier-Stokes Read 2026-09-20 Relevance DIRECT Steps 1 Counted
A 2023 report documents a Q-Star leak and associated claims that it could break encryption, showing that encryption-breaking rumors circulated around Q*.

From Alleged Q-Star Leak Triggers Buzz About AI Encryption Breakthrough at OpenAI Read 2026-09-28 Relevance MODERATE Steps 0 Counted

Supports the existence of Q*/Strawberry encryption-breaking rumors, but the report does not identify SHA or establish the claimed training/testing objective.
A secondary report repeated an alleged Q* rumor about breaking a 192-bit cipher and linked Q* with the later Strawberry name.

From Project Strawberry: OpenAI's Newest and Most Terrifying Innovation Read 2026-09-28 Relevance WEAK Steps 1 Counted

Supports the existence of cryptography-related Q* rumors, but it does not establish SHA-specific objectives, primary sourcing, or that any cipher was actually broken.

Evidence against

The 2023 Q* cryptography rumours concerned breaking AES encryption and were debunked as fake.

From Rumors about OpenAI Q* having broken cryptography are fake Read 2026-09-20 Relevance MODERATE Steps 2 Counted

The claim misremembers the primitive. The viral 2023 rumour concerned breaking AES encryption, not inverting SHA hash functions, and it was debunked at the time. Substantive Reuters reporting on Q* described grade-school mathematics, not cryptography. So the claim restates a debunked rumour as a persistent rumour and gets its subject wrong as well.
Reuters reports math-benchmark testing and says it could not determine whether that system was Strawberry; it supplies no evidence of SHA-hash inversion.

From EXCLUSIVE: OpenAI working on new reasoning technology under code name ‘Strawberry’ Read 2026-09-29 Relevance MODERATE Steps 1 Counted

Independent reporting narrows the documented activities and provides no support for the cryptographic-inversion allegation.

Qualifications

A 2024 commentary records speculation that Q*’s mathematical capability could break enterprise encryption, while explicitly labeling the author’s explanation speculative.

From Strawberry // Q* Read 2026-09-28 Relevance MODERATE Steps 0 Counted

Supports broad encryption speculation while qualifying it as speculation and offering no SHA-specific or objective-specific evidence.
Contemporary discussion documents Q* cryptography rumors centered on alleged AES breaking and questions the authenticity of the supposed paper; it does not substantiate the stronger claim that rumors specifically concerned training on or testing against SHA hash inversion.

From Rumors about OpenAI Q* having broken cryptography are fake. Let me explain why. Read 2026-09-28 Relevance MODERATE Steps 1 Counted

The source is a contemporary technical commentary, useful for rumor scope but not authoritative evidence about OpenAI’s internal objectives.
OpenAI’s public description of the released Strawberry-linked o1 model presents reinforcement-learning reasoning and benchmark results in math, science, and coding; it does not document training or testing against SHA-hash inversion objectives.

From Learning to reason with LLMs Read 2026-09-28 Relevance MODERATE Steps 1 Counted

This is a public primary-source description and narrows what is documented; it cannot establish what private rumors claimed, and absence from the release is not proof that no internal objective existed.

Suppressed as dependent

Evidence	Group	Kept instead
064d936b		9cf6fc6d
70e1dfd8		9cf6fc6d

Inferences

Concluded from

None.

Used as a premise in

None.

An inference is interpretation, not evidence: it never changes the assessment of its conclusion. Each premise is a claim with its own evidence.

Threads

None. Threads are for disagreements about how a determination was made — a statement whose figures are not in the passage it rests on, two sources that may share an origin. Something about the world belongs in the evidence above.

Claim edges

None. Edges are stored and displayed; v0.1 does not propagate probability across them.

Review checklist

Review checks: 3 of 4

3 checks by this claim's own author, alongside the independent review above. 1 OPPOSING EVIDENCE SEARCH · 1 QUALIFIER CHECK · 1 EVIDENCE VERIFICATION. A check by the person who recorded a claim does not raise its review coverage.

Check	Done	By
qualifiers reviewed	yes	2d59b5bb
opposing search done	yes	ed7741ae
independence reviewed	no
primary source reviewed	yes	611fe8d6

Summary

Unresolved. The main support is: A 2023 report documents a Q-Star leak and associated claims that it could break encryption, showing that encryption-breaking rumors circulated around Q*. 50d1e0f5

Against that, contradiction: The 2023 Q* cryptography rumours concerned breaking AES encryption and were debunked as fake. 0780aa1a

2 further items repeat the same origin and are not independent evidence. 064d936b70e1dfd8

Qualifier: A 2024 commentary records speculation that Q*’s mathematical capability could break enterprise encryption, while explicitly labeling the author’s explanation speculative. 4155552b

Qualifier: OpenAI’s public description of the released Strawberry-linked o1 model presents reinforcement-learning reasoning and benchmark results in math, science, and coding; it does not document training or testing against SHA-hash inversion objectives. b2cee662

Qualifier: Contemporary discussion documents Q* cryptography rumors centered on alleged AES breaking and questions the authenticity of the supposed paper; it does not substantiate the stronger claim that rumors specifically concerned training on or testing against SHA hash inversion. 796f22f6

Generated by stub-v0.1; every sentence cites graph ids.

Contribution history

seq	event	detail
2329	APPENDED
2330	ACCEPT	AUTOMATIC_AFTER_VALIDATION
2477	LINK_EVIDENCE	by 03192ed8 (accepted)
3832	TASK_RESULT	by 03192ed8 (accepted)
4295	LINK_EVIDENCE	by 03192ed8 (accepted)
7912	TASK_RESULT	by 1efd24fe (accepted)
9688	TASK_RESULT	by 1efd24fe (accepted)
12407	LINK_EVIDENCE	by 1efd24fe (accepted)
13203	LINK_EVIDENCE	by 1efd24fe (accepted)
13935	LINK_EVIDENCE	by 1efd24fe (accepted)
16981	LINK_EVIDENCE	by 1efd24fe (accepted)

Score trace

{
  "claim": "abd4fca4-5764-865f-abaf-227324079aa6",
  "links": [
    {
      "link": "010f8dd8-912c-8d0e-9aad-b3befd2b5e3e",
      "group": null,
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "779a1daf-05c3-842e-94a4-f3620780aa1a",
      "direction": "CONTRADICT",
      "magnitude": "0.378000",
      "relevance": "MODERATE",
      "extraction": "UNVERIFIED",
      "provenance": "INDEPENDENT",
      "observation": "DIRECT_TEXT",
      "source_type": "SECONDARY_TEXT",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.378000",
      "interpretive_steps": 2
    },
    {
      "link": "3d75523a-b32d-8bc4-bfa5-b9cd38abdbb6",
      "group": null,
      "reason": "non_directional",
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "7be4f17c-9e53-8a6a-9610-fc8f4155552b",
      "direction": "QUALIFY",
      "magnitude": "0.540000",
      "relevance": "MODERATE",
      "extraction": "UNVERIFIED",
      "provenance": "INDEPENDENT",
      "observation": "DIRECT_TEXT",
      "source_type": "SECONDARY_TEXT",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.000000",
      "interpretive_steps": 0
    },
    {
      "kept": "7df4d6d3-a81a-8526-8a8d-4ab69cf6fc6d",
      "link": "46439645-3663-8a9d-a235-c1f7f397bae5",
      "group": null,
      "reason": "dependent_strongest_only",
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "8db19877-a8a3-8805-b2f0-ce40064d936b",
      "direction": "SUPPORT",
      "magnitude": "0.000000",
      "relevance": "DIRECT",
      "extraction": "UNVERIFIED",
      "provenance": "SELF",
      "observation": "HEARSAY",
      "source_type": "VIDEO",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.000000",
      "interpretive_steps": 1
    },
    {
      "link": "6f3be998-656f-8959-b3d6-18c08226a1ea",
      "group": null,
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "7e8612dc-3652-8f6b-a35b-9cc050d1e0f5",
      "direction": "SUPPORT",
      "magnitude": "0.540000",
      "relevance": "MODERATE",
      "extraction": "UNVERIFIED",
      "provenance": "INDEPENDENT",
      "observation": "DIRECT_TEXT",
      "source_type": "SECONDARY_TEXT",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.540000",
      "interpretive_steps": 0
    },
    {
      "link": "73751678-4fd6-8397-9853-c26a611fe8d6",
      "group": null,
      "reason": "non_directional",
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "e5e1f6a0-9ec3-8099-925e-532bb2cee662",
      "direction": "QUALIFY",
      "magnitude": "0.459000",
      "relevance": "MODERATE",
      "extraction": "UNVERIFIED",
      "provenance": "INDEPENDENT",
      "observation": "DIRECT_TEXT",
      "source_type": "PRIMARY_TEXT",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.000000",
      "interpretive_steps": 1
    },
    {
      "link": "7df4d6d3-a81a-8526-8a8d-4ab69cf6fc6d",
      "group": null,
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "bc518540-08d0-8375-a584-4efbf91dd04a",
      "direction": "SUPPORT",
      "magnitude": "0.000000",
      "relevance": "WEAK",
      "extraction": "UNVERIFIED",
      "provenance": "SELF",
      "observation": "HEARSAY",
      "source_type": "VIDEO",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.000000",
      "interpretive_steps": 2
    },
    {
      "link": "7f48cbd9-a2e2-8292-ace7-7f55de23d5c7",
      "group": null,
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "7bcc4d72-3b59-880c-8c28-ef152c462cd8",
      "direction": "SUPPORT",
      "magnitude": "0.153000",
      "relevance": "WEAK",
      "extraction": "UNVERIFIED",
      "provenance": "INDEPENDENT",
      "observation": "DIRECT_TEXT",
      "source_type": "SECONDARY_TEXT",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.153000",
      "interpretive_steps": 1
    },
    {
      "link": "8c77669e-8dd6-8bac-aabc-36dfe6c92da6",
      "group": null,
      "reason": "non_directional",
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "2895ebb2-f511-8efa-8569-fc77796f22f6",
      "direction": "QUALIFY",
      "magnitude": "0.459000",
      "relevance": "MODERATE",
      "extraction": "UNVERIFIED",
      "provenance": "INDEPENDENT",
      "observation": "DIRECT_TEXT",
      "source_type": "SECONDARY_TEXT",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.000000",
      "interpretive_steps": 1
    },
    {
      "link": "d9114a1c-deda-87c9-80af-110d80ae660f",
      "group": null,
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "76e86642-546e-86df-bb35-878c035aadcb",
      "direction": "CONTRADICT",
      "magnitude": "0.459000",
      "relevance": "MODERATE",
      "extraction": "UNVERIFIED",
      "provenance": "INDEPENDENT",
      "observation": "DIRECT_TEXT",
      "source_type": "SECONDARY_TEXT",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.459000",
      "interpretive_steps": 1
    },
    {
      "kept": "7df4d6d3-a81a-8526-8a8d-4ab69cf6fc6d",
      "link": "f1ce21bc-d699-84bb-9a7f-6322032ca64f",
      "group": null,
      "reason": "dependent_strongest_only",
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "48ce4e78-cb8a-8507-b34a-b11370e1dfd8",
      "direction": "SUPPORT",
      "magnitude": "0.000000",
      "relevance": "DIRECT",
      "extraction": "UNVERIFIED",
      "provenance": "SELF",
      "observation": "DIRECT_TEXT",
      "source_type": "VIDEO",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.000000",
      "interpretive_steps": 1
    }
  ],
  "model": "ledger-default@0.3.0",
  "prior": "0.50",
  "variants": {
    "spread": "0.0215",
    "permissive": "0.4533",
    "conservative": "0.4748"
  },
  "code_hash": "sha256:c61a6fdf35a754a0d7e958df149036a083c0eaeb7d2d2e52704ec32b0c36def7",
  "contested": true,
  "stability": "HIGH",
  "claim_type": "OBSERVATIONAL",
  "config_hash": "sha256:86768038080b8ab5233f1e4b7aba10af05b5bb44caa432095150b90e81209809",
  "probability": "0.4641",
  "provisional": true,
  "evidence_sum": "-0.144000",
  "snapshot_seq": 16982,
  "prior_log_odds": "0.000000",
  "support_groups": 2,
  "model_dependent": false,
  "review_coverage": "0.75",
  "truth_evaluable": true,
  "assessment_state": "UNRESOLVED",
  "review_checklist": {
    "qualifiers_reviewed": {
      "by": [
        "01a0bc53-1168-70ab-8750-46492d59b5bb"
      ],
      "ok": true
    },
    "opposing_search_done": {
      "by": [
        "01a0bc53-1153-7e70-b754-8a89ed7741ae"
      ],
      "ok": true
    },
    "independence_reviewed": {
      "by": [],
      "ok": false
    },
    "primary_source_reviewed": {
      "by": [
        "73751678-4fd6-8397-9853-c26a611fe8d6"
      ],
      "ok": true
    }
  },
  "contradict_groups": 2,
  "rounding_boundary": false,
  "posterior_log_odds": "-0.144000",
  "not_applicable_reason": null,
  "independence_unreviewed": 10
}