# OpenAI gpt-oss-20b As of October 22, 2026, OpenAI gpt-oss-20b ranks #97 of 123 models overall with 38.5 ±13.7 on the LLMBenchmark Index. gpt-oss-20b ranks #76 of 108 in math, its strongest category. OpenAI gpt-oss-20b costs $0.03 per million input tokens and $0.15 per million output tokens. gpt-oss-20b was released on August 5, 2025. Page: https://llmbenchmark.ai/models/gpt-oss-20b Data from October 22, 2026. Every figure below names the board it was read from and the day. - Maker: OpenAI - Released: August 5, 2025 - Price per million input tokens: $0.03 - Price per million output tokens: $0.15 - Context window: 131K tokens - Overall rank: 97 of 123, by the LLMBenchmark Index - LLMBenchmark Index: 38.5 ±13.7 (90% margin, from 4 of 7 inputs) - Artificial Analysis Intelligence Index: 10.0 Benchmark results (62 of 195): - GPQA Diamond: 68.9% (source: GPQA (Vals AI), read October 22, 2026) - AIME: 86.0% (source: AIME (Vals AI), read October 22, 2026) - MMLU Pro: 71.6% (source: MMLU Pro (Vals AI), read October 22, 2026) - AA Intelligence Index: 10.0 (source: Artificial Analysis (via OpenRouter), read October 22, 2026) - Design Arena: websites: 852 Elo (source: Design Arena, read October 22, 2026) - Design Arena: data viz: 933 Elo (source: Design Arena, read October 22, 2026) - LiveCodeBench (Vals AI): 80.4% (source: Vals AI boards, read October 22, 2026) - MATH 500 (Vals AI): 94.2% (source: Vals AI boards, read October 22, 2026) - MGSM (Vals AI): 89.0% (source: Vals AI boards, read October 22, 2026) - LegalBench (Vals AI): 70.8% (source: Vals AI boards, read October 22, 2026) - CaseLaw (Vals AI): 43.8% (source: Vals AI boards, read October 22, 2026) - CorpFin (Vals AI): 53.1% (source: Vals AI boards, read October 22, 2026) - TaxEval (Vals AI): 63.7% (source: Vals AI boards, read October 22, 2026) - MedQA (Vals AI): 82.9% (source: Vals AI boards, read October 22, 2026) - AA Math Index: 89.3 (source: Artificial Analysis, read October 22, 2026) - GPQA Diamond (Artificial Analysis run): 68.8% (source: Artificial Analysis, read October 22, 2026) - Humanity's Last Exam (Artificial Analysis run): 11.0% (source: Artificial Analysis, read October 22, 2026) - MMLU Pro (Artificial Analysis run): 74.8% (source: Artificial Analysis, read October 22, 2026) - LiveCodeBench (Artificial Analysis run): 77.7% (source: Artificial Analysis, read October 22, 2026) - SciCode: 38.9% (source: Artificial Analysis, read October 22, 2026) - Terminal-Bench Hard (Artificial Analysis run): 10.6% (source: Artificial Analysis, read October 22, 2026) - IFBench: 65.1% (source: Artificial Analysis, read October 22, 2026) - Long context reasoning (AA-LCR): 34.7% (source: Artificial Analysis, read October 22, 2026) - τ²-Bench Telecom: 60.2% (source: Artificial Analysis, read October 22, 2026) - AIME 2025 (Artificial Analysis run): 89.3% (source: Artificial Analysis, read October 22, 2026) - Terminal-Bench: 3.4% (source: Epoch AI benchmarking hub, read October 22, 2026) - WeirdML: 40.9% (source: Epoch AI benchmarking hub, read October 22, 2026) - GPQA Diamond (Epoch AI run): 60.8% (source: Epoch AI benchmarking hub, read October 22, 2026) - OTIS Mock AIME: 65.3% (source: Epoch AI benchmarking hub, read October 22, 2026) - DTBench: 68.0% (source: Epoch AI benchmarking hub, read October 22, 2026) - LMCA: 14.5% (source: Epoch AI benchmarking hub, read October 22, 2026) - ALE-Bench: 566.0 (source: ALE-Bench (Sakana AI), read October 22, 2026) - Epoch Capabilities Index: 137.7 (source: Epoch AI benchmarking hub, read October 22, 2026) - LMArena Text: 1288 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: coding: 1307 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: hard prompts: 1274 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: math: 1317 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: instruction following: 1236 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: creative writing: 1201 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: English: 1303 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: non-English: 1269 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: Chinese: 1315 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: German: 1255 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: Japanese: 1244 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: Korean: 1236 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: Spanish: 1267 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: Russian: 1278 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: Polish: 1252 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: multi-turn: 1268 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: longer queries: 1250 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: expert: 1258 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: hard prompts in English: 1292 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: software and IT: 1311 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: mathematical work: 1305 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: science: 1296 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: medicine and healthcare: 1310 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: legal and government: 1280 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: business and finance: 1284 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: writing and language: 1238 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: entertainment and media: 1216 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - Kagi LLM Benchmark: 53.2% (source: Kagi LLM Benchmarking Project, read October 22, 2026) - BrowseComp-Plus: 50.6% (source: BrowseComp-Plus, read October 22, 2026)