Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
OpenAI · Proprietary · rank 3 · bench-align-v5
GPT-6.2 Astra is built for reasoning. It ranks in the top quarter for reasoning (top 1% of 216 scored models) and math (top 1% of 111 scored models), by BenchLM's published scores. It accepts text, images and files and produces text.
GPT-6.2 Astra by OpenAI scores 90.8 overall, costs $8.00 per 1M input tokens and $40.00 per 1M output tokens, and carries a context window of 1.05M tokens — observed 2026-11-21.
Seven axes from the ranking source. A missing axis is drawn as a gap.
Every published price component, including cache reads and writes.
Lifecycle events the source associates with this model, plus what this profile does not claim.
That is not evidence the model has no lifecycle plan — only that this source published none.
Each axis shows the ranking source’s score and, beside it, BenchLM’s separate verified score for that axis. The two are published side by side and are not interchangeable.
Measured values with the date they were observed. Nothing is inferred from a sibling model or a provider claim.
Time to first token: latency to the first answer chunk (Artificial Analysis, via BenchLM). Reasoning models include thinking time, so values can run to tens or hundreds of seconds.
Evidence key: ObservedUnavailable
| Measurement | Value | Observed | Last good | Evidence |
|---|---|---|---|---|
| Time to first token | Unavailable | Unavailable | Unavailable | |
| Throughput | 36 tok/s | 2026-11-21 | 2026-11-21 |
Regional or per-endpoint measurements appear only when the API supplies them; none are modelled here.
| $20.00 |
| Batch cache read / 1M tokens (OpenRouter batch) (base rate) | $0.20 |
| Batch cache write / 1M tokens (OpenRouter batch) (base rate) | $5.00 |
Batch API rate (asynchronous jobs), from OpenRouter's batch listing.
The batch rates above are the base batch rate. OpenRouter's batch listing, the same listing as those rates, publishes higher batch rates for requests whose total prompt is above the thresholds below. The cost simulator prices a batch request whose whole prompt is above a threshold entirely at the highest batch tier it exceeds.
10 conversations per day × 8 messages × 22 active days, 1200 input and 400 output tokens per message, no cache. This uses the same calculator as the cost simulator, so an unavailable applicable rate makes the example unavailable too.
Plans from the model's maker, and the API providers that list this exact model today. Links leave ModelChoice; each provider sets its own prices and limits.
| Benchmark | Value | Rank | Tasks | Difficulty | Provenance |
|---|---|---|---|---|---|
| HealthBench Professional (raw) HealthBench Professional raw score | 66.0 | #6 | 525 physician-authored conversations | Professional clinical tasks | HealthBench Professional: Evaluating Large Language Models on Real Clinician Chats |
| GPQA Graduate-Level Google-Proof Q&A | 97 | #1 | 448 questions | Graduate level | GPQA: A Graduate-Level Google-Proof Q&A Benchmark |
| GPQA-D GPQA Diamond | 97.0 | #1 | Graduate-level science questions | Graduate level | GPQA: A Graduate-Level Google-Proof Q&A Benchmark |
| Artificial Analysis Intelligence Index Artificial Analysis Intelligence Index | 62.1 | #2 | Cross-benchmark intelligence index | Display-only external reference | Artificial Analysis |
| AA-Omniscience Index Artificial Analysis Omniscience Index | 50.0 | #4 | Knowledge questions | Broad factual knowledge | AA-Omniscience: Knowledge and Hallucination Benchmark |
| AA-Omniscience Accuracy Artificial Analysis Omniscience Accuracy | 68.6 | #2 | Knowledge questions | Broad knowledge | Artificial Analysis model benchmarks |
| AA-Omniscience Hallucination Rate Artificial Analysis Omniscience Hallucination Rate | 59.1 | #134 | Knowledge questions | Factuality | Artificial Analysis model benchmarks |
| HealthBench Professional HealthBench Professional | 66.3 | #2 | Clinician chat tasks | Professional clinical workflows | HealthBench Professional: Evaluating Large Language Models on Real Clinician Chats |
| Benchmark | Value | Rank | Tasks | Difficulty | Provenance |
|---|---|---|---|---|---|
| Bug Hunt Bench Bug Hunt Bench | 62.7 | #2 | 105 planted bugs across two production TypeScript repositories | Blind production-repository bug finding and repair | Bug Hunt Bench method, caveats, and definitions |
| FrontierCode 2.0 FrontierCode 2.0 | 44.2 | #4 | 132 private software-engineering tasks | Frontier coding-agent quality | FrontierCode 2.0 |
| PostTrainBench PostTrainBench v1.2 | 43.5 |
| Benchmark | Value | Rank | Tasks | Difficulty | Provenance |
|---|---|---|---|---|---|
| FrontierMath v2 (Tiers 1-3) FrontierMath v2 Tiers 1-3 | 92.632 | #4 | 285 private advanced mathematics problems | From olympiad-plus to early research mathematics | FrontierMath v2 benchmark hub |
| FrontierMath v2 (Tier 4) FrontierMath v2 Tier 4 | 100.000 | #1 | 41 private extreme-difficulty mathematics problems | Research-level mathematics requiring hours or days of expert work | FrontierMath Tier 4 v2 leaderboard |
| Benchmark | Value | Rank | Tasks | Difficulty | Provenance |
|---|---|---|---|---|---|
| ARC-AGI-1 ARC-AGI-1 Semi-Private Evaluation | 98.50 | #1 | Semi-private ARC-AGI-1 evaluation set | Abstract visual reasoning | ARC Prize leaderboard |
| ARC-AGI-2 Abstraction and Reasoning Corpus for AGI v2 | 95.833333 | #1 | Visual pattern completion and abstract reasoning | Expert-level — hardest public reasoning benchmark | ARC-AGI-2: A Harder General Intelligence Benchmark |
| ARC-AGI-3 Abstraction and Reasoning Corpus for AGI v3 | 70.306138 |
| Benchmark | Value | Rank | Tasks | Difficulty | Provenance |
|---|---|---|---|---|---|
| Terminal-Bench-Science 0.1 Terminal-Bench-Science 0.1 | 77.1 | #1 | 70 expert-curated scientific research workflows | Frontier scientific research workflows | Terminal-Bench-Science 0.1 |
| AA Briefcase Artificial Analysis Briefcase | 1644 | #12 | Professional knowledge-work tasks | Professional work | Artificial Analysis Briefcase Benchmark Leaderboard |
| AA AutomationBench Artificial Analysis AutomationBench | 77.2 |
| Benchmark | Value | Rank | Tasks | Difficulty | Provenance |
|---|---|---|---|---|---|
| AA-MMMU-Pro Artificial Analysis MMMU-Pro | 89.5 | #1 | Multimodal academic reasoning | Frontier multimodal | Artificial Analysis MMMU-Pro Benchmark Leaderboard |
| Benchmark | Value | Rank | Tasks | Difficulty | Provenance |
|---|---|---|---|---|---|
| ExploitBench ExploitBench v8-bench | 100 | #1 | V8 exploit synthesis runs | Browser exploitation and cybersecurity | ExploitBench |
| SEC-Bench Pro SEC-Bench Pro | 86.8 | #4 | Security engineering tasks | Advanced cybersecurity | Introducing GPT-5.6 |
| FrontierCyber FrontierCyber | 31.3 |
A second, independent score source. Epoch’s own figures in Epoch’s units, one line per model version Epoch evaluated (effort setting as Epoch labels it). They are not part of this site’s capability scores or rank.
| Benchmark | Variant | Score | Epoch date | Evidence |
|---|---|---|---|---|
| GPQA Diamond Accuracy on graduate-level science multiple-choice questions (best score across scorers). Unit: fraction (0–1); higher is better. | max | 0.9684 ± 0.0123 | 2026-10-27 | |
| FrontierMath Tiers 1-3 (v2) Share of private research-level math problems solved, tiers 1-3. Unit: fraction (0–1); higher is better. | max | 0.9263 ± 0.0155 | 2026-10-27 | |
| FrontierMath Tier 4 (v2) Share of private research-level math problems solved, tier 4. Unit: fraction (0–1); higher is better. | max | 1 ± 0 | 2026-10-27 | |
| SimpleQA Verified Short factual questions answered correctly. Unit: fraction (0–1); higher is better. | max | 0.785 ± 0.013 | 2026-10-27 |
Source: Epoch AI, CC BY 4.0 (licence), benchmark data read once a day.
| #2 |
| Post-training four base models on six benchmarks |
| Autonomous ML engineering |
| PostTrainBench |
| #1 |
| Interactive game-like tasks with hidden rules |
| Frontier agentic reasoning |
| ARC-AGI-3: A New Challenge for Frontier Agentic Intelligence |
| MLCR-AA Medical Long Context Reasoning (MLCR-AA) | 39.4 | #9 | Long, fragmented medical-record reasoning | Long-context medical reasoning | Medical Long Context Reasoning (MLCR-AA) |
| ARC-AGI-3 Provider Adapter ARC-AGI-3 Semi-Private: Provider Adapter harness | 99.80 | #1 | — | — | ARC Prize verified ARC-AGI-3 results |
| #4 |
| Business-process automation tasks |
| Agentic automation |
| Artificial Analysis AutomationBench Benchmark Leaderboard |
| AA EnterpriseOps-Gym Artificial Analysis EnterpriseOps-Gym | 29.9 | #31 | Enterprise operations workflows | Enterprise agent operations | Artificial Analysis EnterpriseOps-Gym Benchmark Leaderboard |
| AA Terminal-Bench 4.0 Artificial Analysis Terminal-Bench v4.0 | 64.1 | #3 | Terminal-based agent tasks | Agentic software engineering | Artificial Analysis Terminal-Bench v4.0 Benchmark Leaderboard |
| GDP.pdf Artificial Analysis GDP.pdf | 29.8 | #7 | Professional document-production tasks | Professional knowledge work | GDP.pdf Benchmark Leaderboard |
| AA-AnalystAgent Artificial Analysis AnalystAgent | 62.5 | #1 | Spreadsheet and document analysis questions | Business and data analysis | AA-AnalystAgent Benchmark Leaderboard |
| Harvey LAB-AA v1.1 Harvey LAB-AA v1.1 | 15.0 | #1 | 120 private legal tasks across 24 practice areas | Professional legal work | Harvey LAB-AA v1.1 Benchmark Leaderboard |
| MazeBench MazeBench: visual spatial reasoning in a 3D open world | 17 | #4 | 100 hidden gems across 256 rooms in one open world | Long-horizon spatial planning, with single puzzles that can take over one hundred moves | MazeBench Results |
| CWE-bench CWE-bench v2 | 66.8 | #6 | Audit and patch C/C++, Go, Java, TypeScript/JavaScript, Python, and Rust repositories; 100 public and 150 held-out tasks | Frontier agent evaluation | CWE-bench v2 leaderboard and methodology |
| Terminal-Bench 4.1 Terminal-Bench 4.1 | 62.10 | #5 | 66 professional computer-work tasks | Frontier autonomous knowledge work | Terminal-Bench 4.1 |
| AA Cyber Index Artificial Analysis Cyber Index | 37.3 | #11 | Three defensive cybersecurity evaluations | Real-world vulnerability discovery and patching | Artificial Analysis Cyber Index |
| CWE-Bench-AA Artificial Analysis CWE-Bench-AA | 65.0 | #3 | Real-world vulnerability patching tasks | Automated vulnerability remediation | CWE-Bench-AA Benchmark Leaderboard |
| DeepsecBench-AA Artificial Analysis DeepsecBench-AA | 46.8 | #2 | Vulnerability discovery in real codebases | Security code review | DeepsecBench-AA Benchmark Leaderboard |
| CyberGym-E2E-AA Artificial Analysis CyberGym-E2E-AA | 0.0 | #17 | 131 end-to-end vulnerability reproduction and patching tasks | Real-world cybersecurity | CyberGym-E2E-AA Benchmark Leaderboard |
| MKA Multimodal Knowledge Agent Benchmark | 52.7 | #6 | 200 multimodal professional tasks across 10 industry verticals | Expert multimodal knowledge work | MKA Benchmark Leaderboard |
| LifeVal LifeVal: everyday life tasks | 86.1 | #1 | 169 everyday life tasks (caregiving, health, education, household administration, personal finance, shopping, government services) | Agentic personal-assistant work | LifeVal |
| 197 dynamic cyber tasks |
| Easy through elite cyber operations |
| FrontierCyber |
| CyScenarioBench success CyScenarioBench Average Success Rate | 63 | #3 | 11 cyber scenarios | Long-horizon cybersecurity | Assessing GPT-5.6 Sol |
| CyScenarioBench solved CyScenarioBench Scenarios Ever Solved | 9 | #1 | 11 cyber scenarios | Long-horizon cybersecurity | Assessing GPT-5.6 Sol |
| SpireClimb SpireClimb (Slay the Spire 1 and 2) | 41.3 | #1 | 60 Slay the Spire 1 runs (Ascension 20, Heart kill) and 75 Slay the Spire 2 runs (Ascension 10) per model | Maximum game difficulty, long-horizon computer use | SpireClimb (Epoch AI) |
| Chess Puzzles Chess Puzzles (Epoch AI) | 78.0 | #3 | 100 programmatically generated chess positions with a single best move | Novel chess positions, single best move | Chess Puzzles (Epoch AI) |
| Mystery Game Puzzles Mystery Game Puzzles (Epoch AI) | 95.0 | #1 | 100 single-best-move positions from an undisclosed game | Unfamiliar game positions, single best move | Mystery Game Puzzles (Epoch AI) |
| Furniture Assembly Furniture Assembly Benchmark (Epoch AI) | 89.2 | #2 | 60 photos of three IKEA builds, some with deliberate assembly mistakes | Spatial reasoning from technical diagrams to real photos | Furniture Assembly Benchmark (Epoch AI) |
| EBR-bench Earthborne Rangers (EBR-bench) | 92.4 | #1 | Ten playthroughs of an Earthborne Rangers campaign segment per sample (eight learning, two scored) | Learning an unfamiliar game over multi-hour playthroughs | Earthborne Rangers (EBR-bench), Epoch AI |
| MirrorCode MirrorCode (Epoch AI and METR) | 67.8 | #8 | 30 reimplementation tasks (15 programs in two languages), three attempts each | Days-long software engineering without source code | MirrorCode (Epoch AI) |
| ExploitBench Internal Port ExploitBench — Internal Port (June–August 2026) | 53.5 | #1 | Recently disclosed V8 and Chrome vulnerabilities (June–August 2026) | Browser exploitation | GPT-6 Cyber System Card |
| ExploitGym Honeypot ExploitGym Honeypot out-of-scope exploitation rate | 0.00 | #5 | Hard ExploitGym tasks with an out-of-scope honeypot service | Cyber-agent scope adherence | GPT-6 Cyber: Advancing the frontier of aligned cybersecurity agents |
| FrontierPhysics End-to-End Physics Research Benchmark for AI Agents | 26.8 | #4 | 56 research projects from 42 physicists across nine fields, three trials each (168 rollouts per setup) | Open-ended research projects that took their authors about 270 days on average | FrontierPhysics: Evaluating Agents for End-to-End Physics Research (arXiv, November 4, 2026) |