{"source":"BenchGecko","url":"https://benchgecko.ai/model/gpt-oss-20b","as_of":"2026-10-05","license":"BenchGecko collected data (prices, provider offers, Gecko Tests) is CC BY 4.0 · https://creativecommons.org/licenses/by/4.0/. Benchmark scores belong to their original publishers (see sources) and are aggregated with attribution. Attribution required: \"Source: BenchGecko\" with a link.","attribution":"Source: BenchGecko · https://benchgecko.ai/model/gpt-oss-20b","cite":"gpt-oss-20b · benchmarks, pricing and providers. BenchGecko, data as of 2026-10-05. https://benchgecko.ai/model/gpt-oss-20b","slug":"gpt-oss-20b","name":"gpt-oss-20b","provider":{"name":"OpenAI","slug":"openai"},"model_type":"text","release_date":"2025-08-05","is_open_source":true,"status":"active","context_window":131072,"description":"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...","benchgecko_score":{"value":45,"rank":171,"of":312,"method":"Normalized average of public benchmark scores","method_url":"https://benchgecko.ai/methodology"},"avg_score":45,"list_price":{"input_usd_per_m":0.018,"output_usd_per_m":0.09,"source":"OpenRouter models API (checked daily)","as_of":"2026-10-05"},"pricing":{"input":0.018,"output":0.09},"providers":[{"provider":"Darkbloom","endpoint":"darkbloom/fp8","quantization":"fp8","context_tokens":131072,"input_usd_per_m":0.018,"output_usd_per_m":0.09,"cache_read_usd_per_m":0.009,"uptime_1d_pct":99.84,"as_of":"2026-10-05"},{"provider":"AkashML","endpoint":"akashml/fp4","quantization":"fp4","context_tokens":131072,"input_usd_per_m":0.02,"output_usd_per_m":0.1,"cache_read_usd_per_m":null,"uptime_1d_pct":99.27,"as_of":"2026-10-05"},{"provider":"DekaLLM","endpoint":"dekallm/bf16","quantization":"bf16","context_tokens":131072,"input_usd_per_m":0.029,"output_usd_per_m":0.14,"cache_read_usd_per_m":0.029,"uptime_1d_pct":99.12,"as_of":"2026-10-05"},{"provider":"CoreWeave","endpoint":"coreweave/fp4","quantization":"fp4","context_tokens":131072,"input_usd_per_m":0.03,"output_usd_per_m":0.13,"cache_read_usd_per_m":0.03,"uptime_1d_pct":99.99,"as_of":"2026-10-05"},{"provider":"DeepInfra","endpoint":"deepinfra/bf16","quantization":"bf16","context_tokens":131072,"input_usd_per_m":0.03,"output_usd_per_m":0.14,"cache_read_usd_per_m":null,"uptime_1d_pct":99.95,"as_of":"2026-10-05"},{"provider":"Parasail","endpoint":"parasail/fp4","quantization":"fp4","context_tokens":131072,"input_usd_per_m":0.03,"output_usd_per_m":0.15,"cache_read_usd_per_m":0.02,"uptime_1d_pct":99.75,"as_of":"2026-10-05"},{"provider":"Novita","endpoint":"novita/fp4","quantization":"fp4","context_tokens":131072,"input_usd_per_m":0.04,"output_usd_per_m":0.15,"cache_read_usd_per_m":null,"uptime_1d_pct":99.92,"as_of":"2026-10-05"},{"provider":"SiliconFlow","endpoint":"siliconflow/fp8","quantization":"fp8","context_tokens":131072,"input_usd_per_m":0.04,"output_usd_per_m":0.18,"cache_read_usd_per_m":null,"uptime_1d_pct":91.54,"as_of":"2026-10-05"},{"provider":"Amazon Bedrock","endpoint":"amazon-bedrock","quantization":null,"context_tokens":131072,"input_usd_per_m":0.07,"output_usd_per_m":0.15,"cache_read_usd_per_m":null,"uptime_1d_pct":99.98,"as_of":"2026-10-05"},{"provider":"Amazon Bedrock","endpoint":"amazon-bedrock/eu-west-1","quantization":null,"context_tokens":131072,"input_usd_per_m":0.07,"output_usd_per_m":0.15,"cache_read_usd_per_m":null,"uptime_1d_pct":100,"as_of":"2026-10-05"},{"provider":"Google","endpoint":"google-vertex/us-central1","quantization":null,"context_tokens":131072,"input_usd_per_m":0.07,"output_usd_per_m":0.25,"cache_read_usd_per_m":null,"uptime_1d_pct":91.32,"as_of":"2026-10-05"},{"provider":"Groq","endpoint":"groq","quantization":null,"context_tokens":131072,"input_usd_per_m":0.075,"output_usd_per_m":0.3,"cache_read_usd_per_m":0.0375,"uptime_1d_pct":97.22,"as_of":"2026-10-05"}],"cheapest_provider":{"provider":"Darkbloom","endpoint":"darkbloom/fp8","quantization":"fp8","context_tokens":131072,"input_usd_per_m":0.018,"output_usd_per_m":0.09,"cache_read_usd_per_m":0.009,"uptime_1d_pct":99.84,"as_of":"2026-10-05"},"providers_source":"OpenRouter endpoints API, one row per provider, refreshed daily","scores":[{"benchmark":"Chatbot Arena Elo — Overall","benchmark_slug":"arena-elo-overall","score":1317.58,"unit":"elo","category":"arena","source":"LMArena","source_url":"https://lmarena.ai/leaderboard","benchmark_url":"https://benchgecko.ai/benchmark/arena-elo-overall"},{"benchmark":"OpenCompass — IFEval","benchmark_slug":"oc-ifeval","score":88.9,"unit":"%","category":"language","source":"OpenCompass","source_url":"https://rank.opencompass.org.cn","benchmark_url":"https://benchgecko.ai/benchmark/oc-ifeval"},{"benchmark":"OpenCompass — AIME2025","benchmark_slug":"oc-aime2025","score":87.9,"unit":"%","category":"math","source":"OpenCompass","source_url":"https://rank.opencompass.org.cn","benchmark_url":"https://benchgecko.ai/benchmark/oc-aime2025"},{"benchmark":"HELM — MMLU-Pro","benchmark_slug":"helm-mmlu-pro","score":74,"unit":"%","category":"knowledge","source":"Stanford HELM","source_url":"https://crfm.stanford.edu/helm/","benchmark_url":"https://benchgecko.ai/benchmark/helm-mmlu-pro"},{"benchmark":"HELM — WildBench","benchmark_slug":"helm-wildbench","score":73.7,"unit":"%","category":"reasoning","source":"Stanford HELM","source_url":"https://crfm.stanford.edu/helm/","benchmark_url":"https://benchgecko.ai/benchmark/helm-wildbench"},{"benchmark":"HELM — IFEval","benchmark_slug":"helm-ifeval","score":73.2,"unit":"%","category":"language","source":"Stanford HELM","source_url":"https://crfm.stanford.edu/helm/","benchmark_url":"https://benchgecko.ai/benchmark/helm-ifeval"},{"benchmark":"OpenCompass — MMLU-Pro","benchmark_slug":"oc-mmlu-pro","score":72.8,"unit":"%","category":"knowledge","source":"OpenCompass","source_url":"https://rank.opencompass.org.cn","benchmark_url":"https://benchgecko.ai/benchmark/oc-mmlu-pro"},{"benchmark":"OpenCompass — GPQA-Diamond","benchmark_slug":"oc-gpqa-diamond","score":68.9,"unit":"%","category":"knowledge","source":"OpenCompass","source_url":"https://rank.opencompass.org.cn","benchmark_url":"https://benchgecko.ai/benchmark/oc-gpqa-diamond"},{"benchmark":"Artificial Analysis · GPQA Diamond","benchmark_slug":"aa-gpqa-diamond","score":68.8,"unit":"%","category":"speed","source":"Artificial Analysis","source_url":"https://artificialanalysis.ai","benchmark_url":"https://benchgecko.ai/benchmark/aa-gpqa-diamond"},{"benchmark":"OpenCompass — LiveCodeBenchV6","benchmark_slug":"oc-livecodebenchv6","score":68.4,"unit":"%","category":"coding","source":"OpenCompass","source_url":"https://rank.opencompass.org.cn","benchmark_url":"https://benchgecko.ai/benchmark/oc-livecodebenchv6"},{"benchmark":"OTIS Mock AIME 2024-2025","benchmark_slug":"otis-mock-aime-2024-2025","score":65.24,"unit":"%","category":"math","source":"Epoch AI Benchmarking Hub","source_url":"https://epoch.ai/data/ai-benchmarking-dashboard","benchmark_url":"https://benchgecko.ai/benchmark/otis-mock-aime-2024-2025"},{"benchmark":"Artificial Analysis · IFBench","benchmark_slug":"aa-ifbench","score":65.1,"unit":"%","category":"speed","source":"Artificial Analysis","source_url":"https://artificialanalysis.ai","benchmark_url":"https://benchgecko.ai/benchmark/aa-ifbench"},{"benchmark":"Artificial Analysis · tau2-Bench Telecom","benchmark_slug":"aa-tau2-bench","score":60.2,"unit":"%","category":"speed","source":"Artificial Analysis","source_url":"https://artificialanalysis.ai","benchmark_url":"https://benchgecko.ai/benchmark/aa-tau2-bench"},{"benchmark":"HELM — GPQA","benchmark_slug":"helm-gpqa","score":59.4,"unit":"%","category":"knowledge","source":"Stanford HELM","source_url":"https://crfm.stanford.edu/helm/","benchmark_url":"https://benchgecko.ai/benchmark/helm-gpqa"},{"benchmark":"HELM — Omni-MATH","benchmark_slug":"helm-omni-math","score":56.5,"unit":"%","category":"math","source":"Stanford HELM","source_url":"https://crfm.stanford.edu/helm/","benchmark_url":"https://benchgecko.ai/benchmark/helm-omni-math"},{"benchmark":"GPQA diamond","benchmark_slug":"gpqa-diamond","score":47.73,"unit":"%","category":"knowledge","source":"Epoch AI Benchmarking Hub","source_url":"https://epoch.ai/data/ai-benchmarking-dashboard","benchmark_url":"https://benchgecko.ai/benchmark/gpqa-diamond"},{"benchmark":"Dtbench","benchmark_slug":"dtbench","score":46.67,"unit":"%","category":"general","source":"Epoch AI Benchmarking Hub","source_url":"https://epoch.ai/data/ai-benchmarking-dashboard","benchmark_url":"https://benchgecko.ai/benchmark/dtbench"},{"benchmark":"WeirdML","benchmark_slug":"weirdml","score":40.93,"unit":"%","category":"coding","source":"Epoch AI Benchmarking Hub","source_url":"https://epoch.ai/data/ai-benchmarking-dashboard","benchmark_url":"https://benchgecko.ai/benchmark/weirdml"},{"benchmark":"Artificial Analysis · SciCode","benchmark_slug":"aa-scicode","score":38.9,"unit":"%","category":"speed","source":"Artificial Analysis","source_url":"https://artificialanalysis.ai","benchmark_url":"https://benchgecko.ai/benchmark/aa-scicode"},{"benchmark":"Artificial Analysis · Long Context Reasoning","benchmark_slug":"aa-long-context-reasoning","score":34.7,"unit":"%","category":"speed","source":"Artificial Analysis","source_url":"https://artificialanalysis.ai","benchmark_url":"https://benchgecko.ai/benchmark/aa-long-context-reasoning"},{"benchmark":"Lmca","benchmark_slug":"lmca","score":17.07,"unit":"%","category":"general","source":"Epoch AI Benchmarking Hub","source_url":"https://epoch.ai/data/ai-benchmarking-dashboard","benchmark_url":"https://benchgecko.ai/benchmark/lmca"},{"benchmark":"OpenCompass — HLE","benchmark_slug":"oc-hle","score":11.6,"unit":"%","category":"knowledge","source":"OpenCompass","source_url":"https://rank.opencompass.org.cn","benchmark_url":"https://benchgecko.ai/benchmark/oc-hle"},{"benchmark":"Artificial Analysis · Humanity's Last Exam","benchmark_slug":"aa-humanitys-last-exam","score":11,"unit":"%","category":"speed","source":"Artificial Analysis","source_url":"https://artificialanalysis.ai","benchmark_url":"https://benchgecko.ai/benchmark/aa-humanitys-last-exam"},{"benchmark":"Artificial Analysis · Terminal-Bench Hard","benchmark_slug":"aa-terminal-bench-hard","score":10.6,"unit":"%","category":"speed","source":"Artificial Analysis","source_url":"https://artificialanalysis.ai","benchmark_url":"https://benchgecko.ai/benchmark/aa-terminal-bench-hard"},{"benchmark":"Artificial Analysis — Quality Index","benchmark_slug":"aa-quality-index","score":8.97,"unit":"index","category":"speed","source":"Artificial Analysis","source_url":"https://artificialanalysis.ai","benchmark_url":"https://benchgecko.ai/benchmark/aa-quality-index"},{"benchmark":"Terminal Bench","benchmark_slug":"terminal-bench","score":3.4,"unit":"%","category":"coding","source":"Epoch AI Benchmarking Hub","source_url":"https://epoch.ai/data/ai-benchmarking-dashboard","benchmark_url":"https://benchgecko.ai/benchmark/terminal-bench"},{"benchmark":"Artificial Analysis · CritPt","benchmark_slug":"aa-critpt","score":1.4,"unit":"%","category":"speed","source":"Artificial Analysis","source_url":"https://artificialanalysis.ai","benchmark_url":"https://benchgecko.ai/benchmark/aa-critpt"},{"benchmark":"Artificial Analysis · GDPval","benchmark_slug":"aa-gdpval","score":0,"unit":"%","category":"speed","source":"Artificial Analysis","source_url":"https://artificialanalysis.ai","benchmark_url":"https://benchgecko.ai/benchmark/aa-gdpval"},{"benchmark":"Chess Puzzles","benchmark_slug":"chess-puzzles","score":0,"unit":"%","category":"knowledge","source":"Epoch AI Benchmarking Hub","source_url":"https://epoch.ai/data/ai-benchmarking-dashboard","benchmark_url":"https://benchgecko.ai/benchmark/chess-puzzles"}],"gecko_tests":null,"links":{"page":"https://benchgecko.ai/model/gpt-oss-20b","json":"https://benchgecko.ai/api/v1/models/gpt-oss-20b","price_history":"https://benchgecko.ai/api/v1/price-history/gpt-oss-20b","provider_prices":"https://benchgecko.ai/pricing/arbitrage/gpt-oss-20b"}}