As of October 20, 2026, OpenAI gpt-oss-20b ranks #155 of 213 models overall with 10.0 on the Artificial Analysis Intelligence Index. Its strongest category is reasoning, where it ranks #146 of 203. OpenAI gpt-oss-20b costs $0.018 per million input tokens and $0.09 per million output tokens. It was released on August 5, 2025.
Its output price ranks 13 of 349 priced models, counting from the cheapest.
gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for...
openai/gpt-oss-20b 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 4 of 6 categories. A place counts only the models measured on that board.
No coding benchmark has measured this model, so there is no index. We do not estimate one from its price, its maker or its siblings. How it is weighted
Not measured on SWE-bench Lite, SWE-bench (full test set), SWE-bench Multimodal, SWE-bench Multilingual, AA Coding Index, SWE-bench Verified (Vals AI), Terminal-Bench 4 (Vals AI), Vibe Code Bench (Vals AI), Code migration (Vals AI), ProgramBench (Vals AI), IOI (Vals AI), Terminal-Bench 2.1 (Vals AI), SWE-bench Multimodal (Vals AI), Vibe Code Bench 1-100 (Vals AI), SkillsBench (Vals AI), CyberBench (Vals AI), DeepSWE, FrontierCode, FrontierSWE, GSO-Bench, MirrorCode, Cybench, LMArena WebDev: React, LMArena WebDev: HTML, LMArena WebDev: from an image
Not measured on AA Agentic Index, Design Arena: agents building web apps, Design Arena: agents building full-stack apps, Design Arena: agents building mobile apps, Design Arena: agents building Android apps, Design Arena: agents building Godot games, Design Arena: agents building games, Design Arena: HTML slides, Design Arena: PowerPoint slides, Design Arena: spreadsheets, LiveBench research, LiveBench browser, LiveBench long agentic, SREBench (Vals AI), Finance Agent (Vals AI), Tax Agent Bench (Vals AI), Legal Research Bench (Vals AI), Harvey's Legal Agent Benchmark (Vals AI), Excel Modeling Benchmark (Vals AI), BioMysteryBench (Vals AI), Terminal-Bench-Science (Vals AI), Vals Science Index, Time Horizon Index (Vals AI), CUA-bench (Vals AI), Robo-bench (Vals AI), APEX-Agents, METR time horizons, Remote Labor Index, GDPval, OSWorld 2.0, BALROG, DeepResearch Bench, PostTrainBench, TheAgentCompany, OSWorld, SpireClimb, Vending-Bench 2, LMArena Search
Not measured on LiveBench reasoning, MysteryMechanism (Vals AI), MMMU Pro (Vals AI), ARC-AGI, ARC-AGI-2, Humanity's Last Exam, SimpleBench, Chess puzzles, Mystery game puzzles, EBR-bench, VPCT, BIG-Bench Hard, LMArena Vision: diagrams, LMArena Vision: homework
Not measured on FrontierMath, tiers 1 to 3, FrontierMath, tier 4, ProofBench, MATH level 5, GSM8K
Not measured on MortgageTax (Vals AI), MedCode (Vals AI), SimpleQA Verified, MMLU, GeoBench, AI2 Reasoning Challenge
Not measured on Design Arena: UI components, Design Arena: game dev, Design Arena: 3D, Design Arena: all code categories, Design Arena: SVG, Design Arena: ASCII art, LiveBench design
Not measured on Vals Index, LiveBench data analysis, Public Benefits Bench (Vals AI), SAGE (Vals AI), MedScribe (Vals AI), Vals Multimodal Index, Fiction.LiveBench, Creative writing (Lech Mazur), CL-bench, Surface Evolver Bench, Furniture assembly, ExploitBench, HellaSwag, WinoGrande, PIQA, CL-bench Life, LMArena Vision, LMArena Text: French, LMArena Vision: English, LMArena Vision: Chinese, LMArena Vision: reading text, LMArena Vision: humor, LMArena Vision: captioning, LMArena Vision: recognizing things, LMArena Vision: creative writing, 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 | Not measured | 20% |
| LMArena WebDevCoding | Not measured | 15% |
| Vals IndexGeneral | Not measured | Not in index |
| GPQA DiamondReasoning | 68.9% | Not in index |
| AIMEMath | 86.0% | Not in index |
| MMLU ProKnowledge | 71.6% | Not in index |
| AA Intelligence IndexIntelligence | 10.0 | Not in index |
| AA Coding IndexCoding | Not measured | Not in index |
| AA Agentic IndexAgentic work | Not measured | Not in index |
| Design Arena: websitesWeb design | 853 | Not in index |
| Design Arena: UI componentsWeb design | Not measured | Not in index |
| Design Arena: game devWeb design | Not measured | Not in index |
| Design Arena: data vizWeb design | 933 | Not in index |
| Design Arena: 3DWeb design | Not measured | Not in index |
| Design Arena: all code categoriesWeb design | Not measured | Not in index |
| Design Arena: SVGWeb design | Not measured | Not in index |
| Design Arena: ASCII artWeb design | Not measured | Not in index |
| Design Arena: agents building web appsAgentic work | Not measured | Not in index |
| Design Arena: agents building full-stack appsAgentic work | Not measured | Not in index |
| Design Arena: agents building mobile appsAgentic work | Not measured | Not in index |
| Design Arena: agents building Android appsAgentic work | Not measured | Not in index |
| Design Arena: agents building Godot gamesAgentic work | Not measured | 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 | Not measured | 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 | Not measured | Not in index |
| LiveBench data analysisGeneral | Not measured | Not in index |
| LiveBench researchAgentic work | Not measured | Not in index |
| LiveBench browserAgentic work | Not measured | Not in index |
| LiveBench long agenticAgentic work | Not measured | Not in index |
| LiveBench designWeb design | Not measured | Not in index |
| LiveCodeBench (Vals AI)Coding | 80.4% | Not in index |
| SWE-bench Verified (Vals AI)Coding | Not measured | Not in index |
| Terminal-Bench 4 (Vals AI)Coding | Not measured | Not in index |
| Vibe Code Bench (Vals AI)Coding | Not measured | Not in index |
| Code migration (Vals AI)Coding | Not measured | Not in index |
| ProgramBench (Vals AI)Coding | Not measured | Not in index |
| SREBench (Vals AI)Agentic work | Not measured | Not in index |
| IOI (Vals AI)Coding | Not measured | Not in index |
| Terminal-Bench 2.1 (Vals AI)Coding | Not measured | Not in index |
| SWE-bench Multimodal (Vals AI)Coding | Not measured | Not in index |
| Vibe Code Bench 1-100 (Vals AI)Coding | Not measured | Not in index |
| SkillsBench (Vals AI)Coding | Not measured | Not in index |
| CyberBench (Vals AI)Coding | Not measured | Not in index |
| Finance Agent (Vals AI)Agentic work | Not measured | Not in index |
| Tax Agent Bench (Vals AI)Agentic work | Not measured | Not in index |
| Legal Research Bench (Vals AI)Agentic work | Not measured | Not in index |
| Harvey's Legal Agent Benchmark (Vals AI)Agentic work | Not measured | Not in index |
| Excel Modeling Benchmark (Vals AI)Agentic work | Not measured | Not in index |
| BioMysteryBench (Vals AI)Agentic work | Not measured | Not in index |
| Terminal-Bench-Science (Vals AI)Agentic work | Not measured | Not in index |
| Vals Science IndexAgentic work | Not measured | Not in index |
| Time Horizon Index (Vals AI)Agentic work | Not measured | Not in index |
| CUA-bench (Vals AI)Agentic work | Not measured | Not in index |
| Robo-bench (Vals AI)Agentic work | Not measured | Not in index |
| Public Benefits Bench (Vals AI)General | Not measured | Not in index |
| MysteryMechanism (Vals AI)Reasoning | Not measured | Not in index |
| MMMU Pro (Vals AI)Reasoning | Not measured | Not in index |
| MATH 500 (Vals AI)Math | 94.2% | Not in index |
| MGSM (Vals AI)Math | 89.0% | Not in index |
| SAGE (Vals AI)General | Not measured | Not in index |
| LegalBench (Vals AI)Knowledge | 70.8% | Not in index |
| CaseLaw (Vals AI)Knowledge | 43.8% | Not in index |
| CorpFin (Vals AI)Knowledge | 53.1% | Not in index |
| TaxEval (Vals AI)Knowledge | 63.7% | Not in index |
| MortgageTax (Vals AI)Knowledge | Not measured | Not in index |
| MedQA (Vals AI)Knowledge | 82.9% | Not in index |
| MedCode (Vals AI)Knowledge | Not measured | Not in index |
| MedScribe (Vals AI)General | Not measured | Not in index |
| Vals Multimodal IndexGeneral | Not measured | Not in index |
| AA Math IndexMath | 89.3 | Not in index |
| GPQA Diamond (Artificial Analysis run)Reasoning | 68.8% | Not in index |
| Humanity's Last Exam (Artificial Analysis run)Reasoning | 11.0% | Not in index |
| MMLU Pro (Artificial Analysis run)Knowledge | 74.8% | Not in index |
| LiveCodeBench (Artificial Analysis run)Coding | 77.7% | Not in index |
| SciCodeCoding | 38.9% | Not in index |
| Terminal-Bench Hard (Artificial Analysis run)Coding | 10.6% | Not in index |
| IFBenchGeneral | 65.1% | Not in index |
| Long context reasoning (AA-LCR)General | 34.7% | Not in index |
| τ²-Bench TelecomAgentic work | 60.2% | Not in index |
| AIME 2025 (Artificial Analysis run)Math | 89.3% | Not in index |
| Terminal-BenchCoding | 3.4% | Not in index |
| DeepSWECoding | Not measured | Not in index |
| FrontierCodeCoding | Not measured | Not in index |
| FrontierSWECoding | Not measured | Not in index |
| GSO-BenchCoding | Not measured | Not in index |
| WeirdMLCoding | 40.9% | Not in index |
| MirrorCodeCoding | Not measured | Not in index |
| APEX-AgentsAgentic work | Not measured | Not in index |
| METR time horizonsAgentic work | Not measured | Not in index |
| Remote Labor IndexAgentic work | Not measured | Not in index |
| GDPvalAgentic work | Not measured | Not in index |
| OSWorld 2.0Agentic work | Not measured | Not in index |
| BALROGAgentic work | Not measured | Not in index |
| DeepResearch BenchAgentic work | Not measured | Not in index |
| PostTrainBenchAgentic work | Not measured | Not in index |
| GPQA Diamond (Epoch AI run)Reasoning | 60.8% | Not in index |
| ARC-AGIReasoning | Not measured | Not in index |
| ARC-AGI-2Reasoning | Not measured | Not in index |
| Humanity's Last ExamReasoning | Not measured | Not in index |
| SimpleBenchReasoning | Not measured | Not in index |
| Chess puzzlesReasoning | Not measured | Not in index |
| Mystery game puzzlesReasoning | Not measured | Not in index |
| EBR-benchReasoning | Not measured | Not in index |
| OTIS Mock AIMEMath | 65.3% | Not in index |
| FrontierMath, tiers 1 to 3Math | Not measured | Not in index |
| FrontierMath, tier 4Math | Not measured | Not in index |
| ProofBenchMath | Not measured | Not in index |
| MATH level 5Math | Not measured | Not in index |
| GSM8KMath | Not measured | Not in index |
| SimpleQA VerifiedKnowledge | Not measured | Not in index |
| MMLUKnowledge | Not measured | Not in index |
| GeoBenchKnowledge | Not measured | Not in index |
| DTBenchGeneral | 68.0% | Not in index |
| LMCAGeneral | 14.5% | 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 | Not measured | 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 | Not measured | 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 | 566.0 | Not in index |
| Vending-Bench 2Agentic work | Not measured | Not in index |
| Epoch Capabilities IndexIntelligence | 137.8 | Not in index |
| LMArena TextGeneral | 1287 | Not in index |
| LMArena Text: codingCoding | 1307 | Not in index |
| LMArena Text: hard promptsReasoning | 1274 | Not in index |
| LMArena Text: mathMath | 1317 | Not in index |
| LMArena Text: instruction followingGeneral | 1236 | Not in index |
| LMArena Text: creative writingGeneral | 1201 | Not in index |
| LMArena VisionGeneral | Not measured | Not in index |
| LMArena Text: EnglishGeneral | 1303 | Not in index |
| LMArena Text: non-EnglishGeneral | 1269 | Not in index |
| LMArena Text: ChineseGeneral | 1315 | Not in index |
| LMArena Text: GermanGeneral | 1255 | Not in index |
| LMArena Text: JapaneseGeneral | 1244 | Not in index |
| LMArena Text: KoreanGeneral | 1236 | Not in index |
| LMArena Text: FrenchGeneral | Not measured | Not in index |
| LMArena Text: SpanishGeneral | 1267 | Not in index |
| LMArena Text: RussianGeneral | 1278 | Not in index |
| LMArena Text: PolishGeneral | 1252 | Not in index |
| LMArena Text: multi-turnGeneral | 1268 | Not in index |
| LMArena Text: longer queriesGeneral | 1250 | Not in index |
| LMArena Text: expertReasoning | 1258 | Not in index |
| LMArena Text: hard prompts in EnglishReasoning | 1292 | Not in index |
| LMArena Text: software and ITCoding | 1311 | Not in index |
| LMArena Text: mathematical workMath | 1305 | Not in index |
| LMArena Text: scienceKnowledge | 1296 | Not in index |
| LMArena Text: medicine and healthcareKnowledge | 1310 | Not in index |
| LMArena Text: legal and governmentKnowledge | 1280 | Not in index |
| LMArena Text: business and financeKnowledge | 1284 | Not in index |
| LMArena Text: writing and languageGeneral | 1238 | Not in index |
| LMArena Text: entertainment and mediaGeneral | 1216 | Not in index |
| LMArena WebDev: ReactCoding | Not measured | Not in index |
| LMArena WebDev: HTMLCoding | Not measured | Not in index |
| LMArena WebDev: from an imageCoding | Not measured | Not in index |
| LMArena Vision: EnglishGeneral | Not measured | Not in index |
| LMArena Vision: ChineseGeneral | Not measured | Not in index |
| LMArena Vision: reading textGeneral | Not measured | Not in index |
| LMArena Vision: diagramsReasoning | Not measured | Not in index |
| LMArena Vision: homeworkReasoning | Not measured | 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 | Not measured | 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.
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 |
|---|---|---|---|
| WeirdML | 40.9% | $0.0108 per run | Median lines of code 166 |
| ALE-Bench | 566.0 | $0.0007 per problem | Tokens per problem 5,951 |
| GPQA Diamond | 68.9% | Not published | Latency, as printed 2m 25s |
| AIME | 86.0% | Not published | Latency, as printed 4m 05s |
| MMLU Pro | 71.6% | Not published | Latency, as printed 30 s |
| Design Arena: websites | 853 Elo | Not published | Battles 627 |
| LiveCodeBench (Vals AI) | 80.4% | Not published | Latency, as printed 1m 49s |
| MATH 500 (Vals AI) | 94.2% | Not published | Latency, as printed 31 s |
| MGSM (Vals AI) | 89.0% | Not published | Latency, as printed 5.8 s |
| LegalBench (Vals AI) | 70.8% | Not published | Latency, as printed 7.2 s |
| CaseLaw (Vals AI) | 43.8% | Not published | Latency, as printed 1m 26s |
| CorpFin (Vals AI) | 53.1% | Not published | Latency, as printed 34 s |
| TaxEval (Vals AI) | 63.7% | Not published | Latency, as printed 49 s |
|---|---|---|---|
| MedQA (Vals AI) | 82.9% | Not published | Latency, as printed 25 s |
| LMArena Text | 1287 Elo | Not published | Votes 10,419 · Interval (±) ±6.4 |
| LMArena Text: coding | 1307 Elo | Not published | Votes 2,112 · Interval (±) ±13.1 |
| LMArena Text: hard prompts | 1274 Elo | Not published | Votes 4,718 · Interval (±) ±9.0 |
| LMArena Text: math | 1317 Elo | Not published | Votes 677 · Interval (±) ±22.4 |
| LMArena Text: instruction following | 1236 Elo | Not published | Votes 2,541 · Interval (±) ±12.2 |
| LMArena Text: creative writing | 1201 Elo | Not published | Votes 1,234 · Interval (±) ±17.8 |
| LMArena Text: English | 1303 Elo | Not published | Votes 4,514 · Interval (±) ±9.1 |
| LMArena Text: non-English | 1269 Elo | Not published | Votes 5,894 · Interval (±) ±8.0 |
| LMArena Text: Chinese | 1315 Elo | Not published | Votes 664 · Interval (±) ±22.8 |
| LMArena Text: German | 1255 Elo | Not published | Votes 217 · Interval (±) ±40.4 |
| LMArena Text: Japanese | 1244 Elo | Not published | Votes 246 · Interval (±) ±38.5 |
| LMArena Text: Korean | 1236 Elo | Not published | Votes 228 · Interval (±) ±40.9 |
| LMArena Text: Spanish | 1267 Elo | Not published | Votes 284 · Interval (±) ±38.1 |
| LMArena Text: Russian | 1278 Elo | Not published | Votes 756 · Interval (±) ±21.1 |
| LMArena Text: Polish | 1252 Elo | Not published | Votes 808 · Interval (±) ±22.0 |
| LMArena Text: multi-turn | 1268 Elo | Not published | Votes 1,777 · Interval (±) ±14.5 |
| LMArena Text: longer queries | 1250 Elo | Not published | Votes 2,104 · Interval (±) ±13.3 |
| LMArena Text: expert | 1258 Elo | Not published | Votes 506 · Interval (±) ±26.9 |
| LMArena Text: hard prompts in English | 1292 Elo | Not published | Votes 2,081 · Interval (±) ±13.3 |
| LMArena Text: software and IT | 1311 Elo | Not published | Votes 3,591 · Interval (±) ±10.1 |
| LMArena Text: mathematical work | 1305 Elo | Not published | Votes 585 · Interval (±) ±24.2 |
| LMArena Text: science | 1296 Elo | Not published | Votes 1,669 · Interval (±) ±14.9 |
| LMArena Text: medicine and healthcare | 1310 Elo | Not published | Votes 564 · Interval (±) ±25.3 |
| LMArena Text: legal and government | 1280 Elo | Not published | Votes 678 · Interval (±) ±23.4 |
| LMArena Text: business and finance | 1284 Elo | Not published | Votes 1,835 · Interval (±) ±14.0 |
| LMArena Text: writing and language | 1238 Elo | Not published | Votes 2,171 · Interval (±) ±13.0 |
| LMArena Text: entertainment and media | 1216 Elo | Not published | Votes 1,730 · Interval (±) ±14.8 |
No comparison page of its own yet: one needs a coding index on both sides, and gpt-oss-20b has none. Compare gpt-oss-20b with any model