GPT-6 Astra

System-card reported configuration; reasoning effort unspecified

Observed named-model performance; helpful-only checkpoint omitted as distinct noncatalog identity.

GPT-6 Astra System Card · Section10.1.1.2.3 · reviewed 2026-09-06

Comparator maximum available thinking/reasoning when reported, otherwise best available result; exact setting not identified by this grid. Official public leaderboard; ranking uses pass@1 with pass@4 tiebreaks. Grid records pass@1 only.

Provider-published launch claim, not independent measurement by Google for every row. Official public leaderboard; ranking uses pass@1 with pass@4 tiebreaks. Grid records pass@1 only. Methodology https://deepmind.google/models/evals-methodology/gemini-4-argon, SHA256 ff1df6bdeddc4c0f48840c09a8a7813b10ae7ad9e053892da68dabf64f09bff7. Preserve source rounding. Methodology says capabilities as of September 2026 and results as of October 2026; review date is source retrieval, not evaluation date.

Comparison limit: Provider-published claim. Independent source results are admitted separately with exact checkpoint, effort and compatible protocol; this grid does not establish a matched comparison.

Google Gemini 4 Argon launch and evaluation methodology · Launch grid row 19: CWE-bench 1 / GPT-6 Astra; methodology Additional Details · reviewed 2026-09-30

Comparator maximum available thinking/reasoning when reported, otherwise best available result; exact setting not identified by this grid. Argon self-computed with mini-swe agent. Astra from the official public leaderboard; Fable and Opus from their system cards. Highest scoring thinking level per Datacurve.

Provider-published launch claim, not independent measurement by Google for every row. Argon self-computed with mini-swe agent. Astra from the official public leaderboard; Fable and Opus from their system cards. Highest scoring thinking level per Datacurve. Methodology https://deepmind.google/models/evals-methodology/gemini-4-argon, SHA256 ff1df6bdeddc4c0f48840c09a8a7813b10ae7ad9e053892da68dabf64f09bff7. Preserve source rounding. Methodology says capabilities as of September 2026 and results as of October 2026; review date is source retrieval, not evaluation date.

Comparison limit: Provider-published claim. Independent source results are admitted separately with exact checkpoint, effort and compatible protocol; this grid does not establish a matched comparison.

Google Gemini 4 Argon launch and evaluation methodology · Launch grid row 5: DeepSWE 1.1 / GPT-6 Astra; methodology Additional Details · reviewed 2026-09-30

Maximum reported across reasoning efforts; OpenAI research environment or API

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-6 Astra: A new generation of intelligence · Coding table / DeepSWE v1.1 / GPT‑6 Astra · reviewed 2026-09-06

Reported reasoning effort low. Complex software-engineering tasks in original real codebases. OpenAI research environment or API; same named metric and evaluation chart. No fallback is reported for this OpenAI configuration.

Provider-published lab self-report. Published chart cost per task $1.5952; source-specific cost does not establish consensus workload efficiency. Competitor results are republished from public reports; the page does not establish independent reproduction by OpenAI.

Introducing GPT-6.1 Sol · Chart: DeepSWE / GPT-6 Astra / low · reviewed 2026-09-29

Reported reasoning effort medium. Complex software-engineering tasks in original real codebases. OpenAI research environment or API; same named metric and evaluation chart. No fallback is reported for this OpenAI configuration.

Provider-published lab self-report. Published chart cost per task $3.0755; source-specific cost does not establish consensus workload efficiency. Competitor results are republished from public reports; the page does not establish independent reproduction by OpenAI.

Introducing GPT-6.1 Sol · Chart: DeepSWE / GPT-6 Astra / medium · reviewed 2026-09-29

Reported reasoning effort high. Complex software-engineering tasks in original real codebases. OpenAI research environment or API; same named metric and evaluation chart. No fallback is reported for this OpenAI configuration.

Provider-published lab self-report. Published chart cost per task $3.9237; source-specific cost does not establish consensus workload efficiency. Competitor results are republished from public reports; the page does not establish independent reproduction by OpenAI.

Introducing GPT-6.1 Sol · Chart: DeepSWE / GPT-6 Astra / high · reviewed 2026-09-29

Reported reasoning effort xhigh. Complex software-engineering tasks in original real codebases. OpenAI research environment or API; same named metric and evaluation chart. No fallback is reported for this OpenAI configuration.

Provider-published lab self-report. Published chart cost per task $4.4291; source-specific cost does not establish consensus workload efficiency. Competitor results are republished from public reports; the page does not establish independent reproduction by OpenAI.

Introducing GPT-6.1 Sol · Chart: DeepSWE / GPT-6 Astra / xhigh · reviewed 2026-09-29

Reported reasoning effort max. Complex software-engineering tasks in original real codebases. OpenAI research environment or API; same named metric and evaluation chart. No fallback is reported for this OpenAI configuration.

Provider-published lab self-report. Published chart cost per task $7.4978; source-specific cost does not establish consensus workload efficiency. Competitor results are republished from public reports; the page does not establish independent reproduction by OpenAI.

Introducing GPT-6.1 Sol · Chart: DeepSWE / GPT-6 Astra / max · reviewed 2026-09-29

Maximum reported across reasoning efforts; OpenAI research environment or API; reduced or absent production safeguards

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-6 Astra: A new generation of intelligence · Cybersecurity table / ExploitBench / GPT‑6 Astra · reviewed 2026-09-06

Maximum reported across reasoning efforts; OpenAI research environment or API; reduced or absent production safeguards; 20 vulnerabilities / 13 Chrome releases

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-6 Astra: A new generation of intelligence · Cybersecurity table / ExploitBench (June-Aug 2026) / GPT‑6 Astra · reviewed 2026-09-06

Maximum reported across reasoning efforts; OpenAI research environment or API; reduced or absent production safeguards; v1 offline environment; no runtime package installation; token capped, no wall-clock cap

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-6 Astra: A new generation of intelligence · Cybersecurity table / ExploitGym / GPT‑6 Astra · reviewed 2026-09-06

Reported reasoning effort low. Share of answers with at least one factual error on de-identified conversations where users flagged an earlier model error; deliberately difficult, not representative of typical usage. OpenAI research environment or API; same named metric and evaluation chart. No fallback is reported for this OpenAI configuration.

Provider-published lab self-report. Published chart cost per task $0.2417; source-specific cost does not establish consensus workload efficiency. Competitor results are republished from public reports; the page does not establish independent reproduction by OpenAI.

Comparison limit: Private user-flagged prompt protocol needs separate review before capability scoring; retained as an exact lower-is-better sourced claim.

Introducing GPT-6.1 Sol · Chart: Factual error rate on difficult prompts (lower is better) / GPT-6 Astra / low · reviewed 2026-09-29

Reported reasoning effort medium. Share of answers with at least one factual error on de-identified conversations where users flagged an earlier model error; deliberately difficult, not representative of typical usage. OpenAI research environment or API; same named metric and evaluation chart. No fallback is reported for this OpenAI configuration.

Provider-published lab self-report. Published chart cost per task $0.3103; source-specific cost does not establish consensus workload efficiency. Competitor results are republished from public reports; the page does not establish independent reproduction by OpenAI.

Comparison limit: Private user-flagged prompt protocol needs separate review before capability scoring; retained as an exact lower-is-better sourced claim.

Introducing GPT-6.1 Sol · Chart: Factual error rate on difficult prompts (lower is better) / GPT-6 Astra / medium · reviewed 2026-09-29

Reported reasoning effort high. Share of answers with at least one factual error on de-identified conversations where users flagged an earlier model error; deliberately difficult, not representative of typical usage. OpenAI research environment or API; same named metric and evaluation chart. No fallback is reported for this OpenAI configuration.

Provider-published lab self-report. Published chart cost per task $0.477; source-specific cost does not establish consensus workload efficiency. Competitor results are republished from public reports; the page does not establish independent reproduction by OpenAI.

Comparison limit: Private user-flagged prompt protocol needs separate review before capability scoring; retained as an exact lower-is-better sourced claim.

Introducing GPT-6.1 Sol · Chart: Factual error rate on difficult prompts (lower is better) / GPT-6 Astra / high · reviewed 2026-09-29

Reported reasoning effort xhigh. Share of answers with at least one factual error on de-identified conversations where users flagged an earlier model error; deliberately difficult, not representative of typical usage. OpenAI research environment or API; same named metric and evaluation chart. No fallback is reported for this OpenAI configuration.

Provider-published lab self-report. Published chart cost per task $0.6032; source-specific cost does not establish consensus workload efficiency. Competitor results are republished from public reports; the page does not establish independent reproduction by OpenAI.

Comparison limit: Private user-flagged prompt protocol needs separate review before capability scoring; retained as an exact lower-is-better sourced claim.

Introducing GPT-6.1 Sol · Chart: Factual error rate on difficult prompts (lower is better) / GPT-6 Astra / xhigh · reviewed 2026-09-29

Reported reasoning effort max. Share of answers with at least one factual error on de-identified conversations where users flagged an earlier model error; deliberately difficult, not representative of typical usage. OpenAI research environment or API; same named metric and evaluation chart. No fallback is reported for this OpenAI configuration.

Provider-published lab self-report. Published chart cost per task $0.7865; source-specific cost does not establish consensus workload efficiency. Competitor results are republished from public reports; the page does not establish independent reproduction by OpenAI.

Comparison limit: Private user-flagged prompt protocol needs separate review before capability scoring; retained as an exact lower-is-better sourced claim.

Introducing GPT-6.1 Sol · Chart: Factual error rate on difficult prompts (lower is better) / GPT-6 Astra / max · reviewed 2026-09-29

Reported reasoning effort max. Adversarial test of whether agents disclose a broken search tool. OpenAI research environment or API; same named metric and evaluation chart. No fallback is reported for this OpenAI configuration.

Provider-published lab self-report. Safety stress outcomes are retained for inspection and excluded from capability scoring. Competitor results are republished from public reports; the page does not establish independent reproduction by OpenAI.

Comparison limit: Safety failure behavior is not a capability or intelligence score.

Introducing GPT-6.1 Sol · Chart: Failure to disclose a broken search tool (lower is better) / GPT-6 Astra / max · reviewed 2026-09-29

Maximum reported across reasoning efforts; OpenAI research environment or API; Codex-style developer instruction on tests, reuse and repository conventions

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-6 Astra: A new generation of intelligence · Coding table / FrontierCode 1.1 Extended (score) / GPT‑6 Astra · reviewed 2026-09-06

Maximum reported across reasoning efforts; OpenAI research environment or API; Codex-style developer instruction on tests, reuse and repository conventions

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-6 Astra: A new generation of intelligence · Coding table / FrontierCode 1.1 Main (score) / GPT‑6 Astra · reviewed 2026-09-06

Maximum reported across reasoning efforts; OpenAI research environment or API

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-6 Astra: A new generation of intelligence · Academic table / FrontierMath Tier 4 (v2) / GPT‑6 Astra · reviewed 2026-09-06