As of October 20, 2026, OpenAI GPT-6 Astra ranks #4 of 213 models overall with 52.7 on the Artificial Analysis Intelligence Index. Its strongest category is math, where it ranks #1 of 116. OpenAI GPT-6 Astra costs $10.00 per million input tokens and $50.00 per million output tokens. It was released on September 4, 2026.
Its output price ranks 337 of 349 priced models, counting from the cheapest.
GPT-6 Astra is OpenAI's flagship model for demanding end-to-end work. It is suited for advanced analysis, software engineering, deep research, scientific work, and document creation, with particular strengths in long-horizon...
openai/gpt-6-astra The tick on a coding bar is the median of every model measured on it. Outside coding, the place is this model's rank on that board. A board that never ran the model says so.
Measured in 5 of 6 categories. A place counts only the models measured on that board.
The index averages only the coding benchmarks that have a result; the others are skipped, never filled in. How it is weighted
Not measured on SWE-bench Lite, SWE-bench (full test set), SWE-bench Multimodal, SWE-bench Multilingual, LiveCodeBench (Vals AI), SWE-bench Verified (Vals AI), SkillsBench (Vals AI), LiveCodeBench (Artificial Analysis run), Terminal-Bench Hard (Artificial Analysis run), Terminal-Bench, Cybench, LMArena Text: coding, LMArena Text: software and IT
Not measured on Design Arena: agents building Android apps, Design Arena: agents building games, Design Arena: HTML slides, Design Arena: spreadsheets, τ²-Bench Telecom, METR time horizons, GDPval, OSWorld 2.0, DeepResearch Bench, PostTrainBench, TheAgentCompany, OSWorld, LMArena Search
Not measured on GPQA Diamond, MMMU Pro (Vals AI), SimpleBench, VPCT, BIG-Bench Hard, LMArena Text: hard prompts, LMArena Text: expert, LMArena Text: hard prompts in English
Not measured on AIME, MATH 500 (Vals AI), MGSM (Vals AI), AA Math Index, AIME 2025 (Artificial Analysis run), MATH level 5, GSM8K, LMArena Text: math, LMArena Text: mathematical work
Not measured on MMLU Pro, LegalBench (Vals AI), CaseLaw (Vals AI), CorpFin (Vals AI), TaxEval (Vals AI), MortgageTax (Vals AI), MedQA (Vals AI), MMLU Pro (Artificial Analysis run), MMLU, GeoBench, AI2 Reasoning Challenge, LMArena Text: science, LMArena Text: medicine and healthcare, LMArena Text: legal and government, LMArena Text: business and finance
Not measured on Public Benefits Bench (Vals AI), Vals Multimodal Index, IFBench, Fiction.LiveBench, Creative writing (Lech Mazur), CL-bench, Surface Evolver Bench, ExploitBench, HellaSwag, WinoGrande, PIQA, CL-bench Life, LMArena Text, LMArena Text: instruction following, LMArena Text: creative writing, LMArena Text: English, LMArena Text: non-English, LMArena Text: Chinese, LMArena Text: German, LMArena Text: Japanese, LMArena Text: Korean, LMArena Text: French, LMArena Text: Spanish, LMArena Text: Russian, LMArena Text: Polish, LMArena Text: multi-turn, LMArena Text: longer queries, LMArena Text: writing and language, LMArena Text: entertainment and media, LMArena Vision: humor, LMArena Vision: captioning, LMArena Vision: recognizing things, LMArena Document
Not measured on Design Arena: images, Design Arena: graphic design, Design Arena: logos, LMArena Text to Image, LMArena Image Edit
Public boards shown as published (Vals AI, Artificial Analysis, Design Arena) and kept out of the coding index. Each links to its source on the benchmark page.
| Benchmark | Result | Index weight |
|---|---|---|
| Aider polyglotCoding | Not measured | 35% |
| SWE-bench VerifiedCoding | Not measured | 30% |
| SWE-bench LiteCoding | Not measured | Not in index |
| SWE-bench (full test set)Coding | Not measured | Not in index |
| SWE-bench MultimodalCoding | Not measured | Not in index |
| SWE-bench MultilingualCoding | Not measured | Not in index |
| LiveBench codingCoding | 93.7% | 20% |
| LMArena WebDevCoding | 1774 | 15% |
| Vals IndexGeneral | 65.0% | Not in index |
| GPQA DiamondReasoning | Not measured | Not in index |
| AIMEMath | Not measured | Not in index |
| MMLU ProKnowledge | Not measured | Not in index |
| AA Intelligence IndexIntelligence | 52.7 | Not in index |
| AA Coding IndexCoding | 76.9 | Not in index |
| AA Agentic IndexAgentic work | 51.0 | Not in index |
| Design Arena: websitesWeb design | 1338 | Not in index |
| Design Arena: UI componentsWeb design | 1400 | Not in index |
| Design Arena: game devWeb design | 1474 | Not in index |
| Design Arena: data vizWeb design | 1333 | Not in index |
| Design Arena: 3DWeb design | 1505 | Not in index |
| Design Arena: all code categoriesWeb design | 1379 | Not in index |
| Design Arena: SVGWeb design | 1431 | Not in index |
| Design Arena: ASCII artWeb design | 1356 | Not in index |
| Design Arena: agents building web appsAgentic work | 1373 | Not in index |
| Design Arena: agents building full-stack appsAgentic work | 1357 | Not in index |
| Design Arena: agents building mobile appsAgentic work | 1308 | Not in index |
| Design Arena: agents building Android appsAgentic work | Not measured | Not in index |
| Design Arena: agents building Godot gamesAgentic work | 1348 | Not in index |
| Design Arena: agents building gamesAgentic work | Not measured | Not in index |
| Design Arena: HTML slidesAgentic work | Not measured | Not in index |
| Design Arena: PowerPoint slidesAgentic work | 1245 | Not in index |
| Design Arena: spreadsheetsAgentic work | Not measured | Not in index |
| Design Arena: imagesImage generation | Not measured | Not in index |
| Design Arena: graphic designImage generation | Not measured | Not in index |
| Design Arena: logosImage generation | Not measured | Not in index |
| LiveBench reasoningReasoning | 92.7% | Not in index |
| LiveBench data analysisGeneral | 83.0% | Not in index |
| LiveBench researchAgentic work | 85.2% | Not in index |
| LiveBench browserAgentic work | 84.0% | Not in index |
| LiveBench long agenticAgentic work | 82.4% | Not in index |
| LiveBench designWeb design | 82.0% | Not in index |
| LiveCodeBench (Vals AI)Coding | Not measured | Not in index |
| SWE-bench Verified (Vals AI)Coding | Not measured | Not in index |
| Terminal-Bench 4 (Vals AI)Coding | 59.6% | Not in index |
| Vibe Code Bench (Vals AI)Coding | 89.6% | Not in index |
| Code migration (Vals AI)Coding | 67.7% | Not in index |
| ProgramBench (Vals AI)Coding | 5.5% | Not in index |
| SREBench (Vals AI)Agentic work | 56.9% | Not in index |
| IOI (Vals AI)Coding | 100.0% | Not in index |
| Terminal-Bench 2.1 (Vals AI)Coding | 87.3% | Not in index |
| SWE-bench Multimodal (Vals AI)Coding | 43.3% | Not in index |
| Vibe Code Bench 1-100 (Vals AI)Coding | 27.6% | Not in index |
| SkillsBench (Vals AI)Coding | Not measured | Not in index |
| CyberBench (Vals AI)Coding | 41.1% | Not in index |
| Finance Agent (Vals AI)Agentic work | 53.5% | Not in index |
| Tax Agent Bench (Vals AI)Agentic work | 20.7% | Not in index |
| Legal Research Bench (Vals AI)Agentic work | 39.4% | Not in index |
| Harvey's Legal Agent Benchmark (Vals AI)Agentic work | 7.1% | Not in index |
| Excel Modeling Benchmark (Vals AI)Agentic work | 71.7% | Not in index |
| BioMysteryBench (Vals AI)Agentic work | 79.3% | Not in index |
| Terminal-Bench-Science (Vals AI)Agentic work | 62.9% | Not in index |
| Vals Science IndexAgentic work | 65.1% | Not in index |
| Time Horizon Index (Vals AI)Agentic work | 90.5% | Not in index |
| CUA-bench (Vals AI)Agentic work | 19.2% | Not in index |
| Robo-bench (Vals AI)Agentic work | 23.3% | Not in index |
| Public Benefits Bench (Vals AI)General | Not measured | Not in index |
| MysteryMechanism (Vals AI)Reasoning | 53.1% | Not in index |
| MMMU Pro (Vals AI)Reasoning | Not measured | Not in index |
| MATH 500 (Vals AI)Math | Not measured | Not in index |
| MGSM (Vals AI)Math | Not measured | Not in index |
| SAGE (Vals AI)General | 46.4% | Not in index |
| LegalBench (Vals AI)Knowledge | Not measured | Not in index |
| CaseLaw (Vals AI)Knowledge | Not measured | Not in index |
| CorpFin (Vals AI)Knowledge | Not measured | Not in index |
| TaxEval (Vals AI)Knowledge | Not measured | Not in index |
| MortgageTax (Vals AI)Knowledge | Not measured | Not in index |
| MedQA (Vals AI)Knowledge | Not measured | Not in index |
| MedCode (Vals AI)Knowledge | 48.5% | Not in index |
| MedScribe (Vals AI)General | 87.9% | Not in index |
| Vals Multimodal IndexGeneral | Not measured | Not in index |
| AA Math IndexMath | Not measured | Not in index |
| GPQA Diamond (Artificial Analysis run)Reasoning | 96.3% | Not in index |
| Humanity's Last Exam (Artificial Analysis run)Reasoning | 54.7% | Not in index |
| MMLU Pro (Artificial Analysis run)Knowledge | Not measured | Not in index |
| LiveCodeBench (Artificial Analysis run)Coding | Not measured | Not in index |
| SciCodeCoding | 56.5% | Not in index |
| Terminal-Bench Hard (Artificial Analysis run)Coding | Not measured | Not in index |
| IFBenchGeneral | Not measured | Not in index |
| Long context reasoning (AA-LCR)General | 80.7% | Not in index |
| τ²-Bench TelecomAgentic work | Not measured | Not in index |
| AIME 2025 (Artificial Analysis run)Math | Not measured | Not in index |
| Terminal-BenchCoding | Not measured | Not in index |
| DeepSWECoding | 74.1% | Not in index |
| FrontierCodeCoding | 53.3% | Not in index |
| FrontierSWECoding | 65.5% | Not in index |
| GSO-BenchCoding | 79.4% | Not in index |
| WeirdMLCoding | 93.3% | Not in index |
| MirrorCodeCoding | 46.7% | Not in index |
| APEX-AgentsAgentic work | 64.7% | Not in index |
| METR time horizonsAgentic work | Not measured | Not in index |
| Remote Labor IndexAgentic work | 20.8% | Not in index |
| GDPvalAgentic work | Not measured | Not in index |
| OSWorld 2.0Agentic work | Not measured | Not in index |
| BALROGAgentic work | 68.3% | Not in index |
| DeepResearch BenchAgentic work | Not measured | Not in index |
| PostTrainBenchAgentic work | Not measured | Not in index |
| GPQA Diamond (Epoch AI run)Reasoning | 95.8% | Not in index |
| ARC-AGIReasoning | 98.5% | Not in index |
| ARC-AGI-2Reasoning | 95.0% | Not in index |
| Humanity's Last ExamReasoning | 54.8% | Not in index |
| SimpleBenchReasoning | Not measured | Not in index |
| Chess puzzlesReasoning | 74.0% | Not in index |
| Mystery game puzzlesReasoning | 84.0% | Not in index |
| EBR-benchReasoning | 76.2% | Not in index |
| OTIS Mock AIMEMath | 100.0% | Not in index |
| FrontierMath, tiers 1 to 3Math | 93.7% | Not in index |
| FrontierMath, tier 4Math | 97.6% | Not in index |
| ProofBenchMath | 99.0% | Not in index |
| MATH level 5Math | Not measured | Not in index |
| GSM8KMath | Not measured | Not in index |
| SimpleQA VerifiedKnowledge | 75.6% | Not in index |
| MMLUKnowledge | Not measured | Not in index |
| GeoBenchKnowledge | Not measured | Not in index |
| DTBenchGeneral | 97.3% | Not in index |
| LMCAGeneral | 64.4% | Not in index |
| Fiction.LiveBenchGeneral | Not measured | Not in index |
| Creative writing (Lech Mazur)General | Not measured | Not in index |
| CL-benchGeneral | Not measured | Not in index |
| Surface Evolver BenchGeneral | Not measured | Not in index |
| Furniture assemblyGeneral | 80.0% | Not in index |
| ExploitBenchGeneral | Not measured | Not in index |
| CybenchCoding | Not measured | Not in index |
| TheAgentCompanyAgentic work | Not measured | Not in index |
| OSWorldAgentic work | Not measured | Not in index |
| SpireClimbAgentic work | 25.7% | Not in index |
| VPCTReasoning | Not measured | Not in index |
| BIG-Bench HardReasoning | Not measured | Not in index |
| AI2 Reasoning ChallengeKnowledge | Not measured | Not in index |
| HellaSwagGeneral | Not measured | Not in index |
| WinoGrandeGeneral | Not measured | Not in index |
| PIQAGeneral | Not measured | Not in index |
| CL-bench LifeGeneral | Not measured | Not in index |
| ALE-BenchCoding | 2951.3 | Not in index |
| Vending-Bench 2Agentic work | $15,515 | Not in index |
| Epoch Capabilities IndexIntelligence | 166.4 | Not in index |
| LMArena TextGeneral | Not measured | Not in index |
| LMArena Text: codingCoding | Not measured | Not in index |
| LMArena Text: hard promptsReasoning | Not measured | Not in index |
| LMArena Text: mathMath | Not measured | Not in index |
| LMArena Text: instruction followingGeneral | Not measured | Not in index |
| LMArena Text: creative writingGeneral | Not measured | Not in index |
| LMArena VisionGeneral | 1280 | Not in index |
| LMArena Text: EnglishGeneral | Not measured | Not in index |
| LMArena Text: non-EnglishGeneral | Not measured | Not in index |
| LMArena Text: ChineseGeneral | Not measured | Not in index |
| LMArena Text: GermanGeneral | Not measured | Not in index |
| LMArena Text: JapaneseGeneral | Not measured | Not in index |
| LMArena Text: KoreanGeneral | Not measured | Not in index |
| LMArena Text: FrenchGeneral | Not measured | Not in index |
| LMArena Text: SpanishGeneral | Not measured | Not in index |
| LMArena Text: RussianGeneral | Not measured | Not in index |
| LMArena Text: PolishGeneral | Not measured | Not in index |
| LMArena Text: multi-turnGeneral | Not measured | Not in index |
| LMArena Text: longer queriesGeneral | Not measured | Not in index |
| LMArena Text: expertReasoning | Not measured | Not in index |
| LMArena Text: hard prompts in EnglishReasoning | Not measured | Not in index |
| LMArena Text: software and ITCoding | Not measured | Not in index |
| LMArena Text: mathematical workMath | Not measured | Not in index |
| LMArena Text: scienceKnowledge | Not measured | Not in index |
| LMArena Text: medicine and healthcareKnowledge | Not measured | Not in index |
| LMArena Text: legal and governmentKnowledge | Not measured | Not in index |
| LMArena Text: business and financeKnowledge | Not measured | Not in index |
| LMArena Text: writing and languageGeneral | Not measured | Not in index |
| LMArena Text: entertainment and mediaGeneral | Not measured | Not in index |
| LMArena WebDev: ReactCoding | 1798 | Not in index |
| LMArena WebDev: HTMLCoding | 1775 | Not in index |
| LMArena WebDev: from an imageCoding | 1726 | Not in index |
| LMArena Vision: EnglishGeneral | 1280 | Not in index |
| LMArena Vision: ChineseGeneral | 1361 | Not in index |
| LMArena Vision: reading textGeneral | 1292 | Not in index |
| LMArena Vision: diagramsReasoning | 1297 | Not in index |
| LMArena Vision: homeworkReasoning | 1337 | Not in index |
| LMArena Vision: humorGeneral | Not measured | Not in index |
| LMArena Vision: captioningGeneral | Not measured | Not in index |
| LMArena Vision: recognizing thingsGeneral | Not measured | Not in index |
| LMArena Vision: creative writingGeneral | 1286 | Not in index |
| LMArena SearchAgentic work | Not measured | Not in index |
| LMArena DocumentGeneral | Not measured | Not in index |
| LMArena Text to ImageImage generation | Not measured | Not in index |
| LMArena Image EditImage generation | Not measured | Not in index |
A list price says what a token costs. The cost per solved task says what a working answer costs, which is the number that decides.
Neither board that publishes what a SWE-bench Verified attempt costs, Vals AI and swebench.com, has run this model. We leave the number out rather than estimate it.
LiveBench, on its own coding tasks, publishes what one task cost this model: $3.315, at a score of 93.7%. LiveBench counts that as $3.538 per successful task. Different tasks from SWE-bench, so it is not our number and does not enter its ranking. LiveBench coding
Beside a result, many boards publish what the run cost, how long it took, the tokens it spent or the votes behind a rating. Each figure as its board printed it, one board per row.
| Board | Result | What the run cost | Also published |
|---|---|---|---|
| LiveBench coding | 93.7% | $3.3154 per task | Output tokens per task 13,904 |
| Vals Index | 65.0% | $21.28 per test | Duration, as printed 33m 39s |
| LiveBench reasoning | 92.7% | $0.3898 per task | Output tokens per task 7,577 |
| LiveBench data analysis | 83.0% | $0.6989 per task | Output tokens per task 13,240 |
| LiveBench research | 85.2% | $17.6599 per task | Output tokens per task 33,608 |
| LiveBench browser | 84.0% | $5.9097 per task | Output tokens per task 3,639 |
| LiveBench long agentic | 82.4% | $38.3789 per task | Output tokens per task 210,429 |
| LiveBench design | 82.0% | $2.8756 per task | Output tokens per task 24,627 |
| Terminal-Bench 4 (Vals AI) | 59.6% | $9.58 per test | Duration, as printed 35m 42s |
| Vibe Code Bench (Vals AI) | 89.6% | $38.51 per test | Duration, as printed 39m 04s |
| Code migration (Vals AI) | 67.7% | $44.36 per test | Duration, as printed 58m 30s |
| ProgramBench (Vals AI) | 5.5% | $11.57 per test | Duration, as printed 27m 02s |
| SREBench (Vals AI) | 56.9% | $10.24 per test | Duration, as printed 17m 51s |
|---|---|---|---|
| IOI (Vals AI) | 100.0% | $6.50 per test | Duration, as printed 31m 40s |
| Terminal-Bench 2.1 (Vals AI) | 87.3% | $1.34 per test | Duration, as printed 4m 34s |
| SWE-bench Multimodal (Vals AI) | 43.3% | $1.81 per test | Duration, as printed 7m 03s |
| Vibe Code Bench 1-100 (Vals AI) | 27.6% | $84.72 per test | Duration, as printed 3h 26m |
| CyberBench (Vals AI) | 41.1% | $1.84 per test | Duration, as printed 4m 10s |
| Finance Agent (Vals AI) | 53.5% | $6.82 per test | Duration, as printed 11m 33s |
| Tax Agent Bench (Vals AI) | 20.7% | $5.80 per test | Duration, as printed 15m 01s |
| Legal Research Bench (Vals AI) | 39.4% | $10.58 per test | Duration, as printed 24m 51s |
| Harvey's Legal Agent Benchmark (Vals AI) | 7.1% | $25.46 per test | Duration, as printed 32m 33s |
| Excel Modeling Benchmark (Vals AI) | 71.7% | $5.83 per test | Duration, as printed 13m 24s |
| BioMysteryBench (Vals AI) | 79.3% | $1.72 per test | Duration, as printed 4m 46s |
| Terminal-Bench-Science (Vals AI) | 62.9% | $20.80 per test | Duration, as printed 1h 49m |
| Vals Science Index | 65.1% | $8.03 per test | Duration, as printed 40m 37s |
| MysteryMechanism (Vals AI) | 53.1% | $1.56 per test | Duration, as printed 8m 09s |
| DeepSWE | 74.1% | $4.4291 mean cost per task | Solved in 4 tries 80.5% · Output tokens per task 29,557 · Agent steps per task 29 |
| FrontierSWE | 65.5% | $1029.65 average cost per task | Best of 5 runs 73.0% · Hours per task 12.1 h |
| WeirdML | 93.3% | $7.7485 per run | Median lines of code 485 |
| ProofBench | 99.0% | $1.7124 per test | Time per proof 5m 08s |
| ALE-Bench | 2951.3 | $2.8641 per problem | Tokens per problem 60,104 |
| LMArena WebDev | 1774 Elo | Not published | Votes 6,169 · Interval (±) ±10.2 |
| Design Arena: websites | 1338 Elo | Not published | Battles 2,236 |
| Design Arena: UI components | 1400 Elo | Not published | Battles 561 |
| Design Arena: game dev | 1474 Elo | Not published | Battles 631 |
| Design Arena: data viz | 1333 Elo | Not published | Battles 356 |
| Design Arena: 3D | 1505 Elo | Not published | Battles 251 |
| Design Arena: all code categories | 1379 Elo | Not published | Battles 8,813 |
| Design Arena: SVG | 1431 Elo | Not published | Battles 1,626 |
| Design Arena: ASCII art | 1356 Elo | Not published | Battles 167 |
| Design Arena: agents building web apps | 1373 Elo | Not published | Battles 650 |
| Design Arena: agents building full-stack apps | 1357 Elo | Not published | Battles 553 |
| Design Arena: agents building mobile apps | 1308 Elo | Not published | Battles 588 |
| Design Arena: agents building Godot games | 1348 Elo | Not published | Battles 328 |
| Design Arena: PowerPoint slides | 1245 Elo | Not published | Battles 804 |
| SAGE (Vals AI) | 46.4% | Not published | Latency, as printed 42 s |
| MedCode (Vals AI) | 48.5% | Not published | Latency, as printed 1m 43s |
| MedScribe (Vals AI) | 87.9% | Not published | Latency, as printed 2m 34s |
| APEX-Agents | 64.7% | Not published | Mean score 75.1% |
| LMArena Vision | 1280 Elo | Not published | Votes 4,012 · Interval (±) ±10.3 |
| LMArena WebDev: React | 1798 Elo | Not published | Votes 3,665 · Interval (±) ±13.0 |
| LMArena WebDev: HTML | 1775 Elo | Not published | Votes 1,205 · Interval (±) ±22.1 |
| LMArena WebDev: from an image | 1726 Elo | Not published | Votes 3,181 · Interval (±) ±12.4 |
| LMArena Vision: English | 1280 Elo | Not published | Votes 1,343 · Interval (±) ±16.9 |
| LMArena Vision: Chinese | 1361 Elo | Not published | Votes 259 · Interval (±) ±37.0 |
| LMArena Vision: reading text | 1292 Elo | Not published | Votes 2,740 · Interval (±) ±11.8 |
| LMArena Vision: diagrams | 1297 Elo | Not published | Votes 966 · Interval (±) ±19.5 |
| LMArena Vision: homework | 1337 Elo | Not published | Votes 560 · Interval (±) ±24.8 |
| LMArena Vision: creative writing | 1286 Elo | Not published | Votes 408 · Interval (±) ±30.8 |