gpt-oss-20b pricing & benchmarks
gpt-oss-20b is available from OpenAI at $0.018 per million input tokens and $0.090 per million output tokens ($0.036 blended at 3:1). It accepts up to 131,072 tokens of context. 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...
Pricing
Per million tokens · OpenRouter
Input$0.018
Output$0.090
Blended (3:1)$0.036
Specification
openai/gpt-oss-20b
Context window131K
Max output33K
ReleasedAug 2025
LicenceOpen weights
Modalitiesin:text, out:text
Benchmarks
Independently run — not vendor self-reports. Every score links its source.
| Benchmark | Score | Config | Run | Source |
|---|---|---|---|---|
| GPQA Diamond | 60.8% ±2.5 | medium effort | 2026-08-27 | Independently run · Epoch AI |
| OTIS Mock AIME | 65.3% ±5.9 | medium effort | 2026-08-27 | Independently run · Epoch AI |
Scores published by Epoch AI, 'AI Benchmarking Hub', used under CC BY 4.0.
Where to run gpt-oss-20b
11 options across 10 providers · 3.6× between cheapest and dearest. Blended at 3:1. ⚠ marks an option serving less context than the best available, or re-pricing above a prompt-length threshold. Bracketed labels are the provider's own service tier or region.
| Host | Blended $/1M | In / Out | Context | Weights | Uptime 24h |
|---|---|---|---|---|---|
| Darkbloom | $0.036 | $0.018 / $0.090 | 131K | fp8 | 100.0% |
| AkashML | $0.040 | $0.020 / $0.100 | 131K | fp4 | 100.0% |
| CoreWeave | $0.055 | $0.030 / $0.130 | 131K | fp4 | 100.0% |
| DekaLLM | $0.057 | $0.029 / $0.140 | 131K | bf16 | 100.0% |
| DeepInfra | $0.058 | $0.030 / $0.140 | 131K | bf16 | 99.0% |
| Parasail | $0.060 | $0.030 / $0.150 | 131K | fp4 | 99.5% |
| SiliconFlow | $0.075 | $0.040 / $0.180 | 131K | fp8 | 91.3% |
| Amazon Bedrock (eu-west-1) | $0.090 | $0.070 / $0.150 | 131K | — | 100.0% |
| Amazon Bedrock | $0.090 | $0.070 / $0.150 | 131K | — | 100.0% |
| Google (us-central1) | $0.115 | $0.070 / $0.250 | 131K | — | 69.6% |
| Groq | $0.131 | $0.075 / $0.300 | 131K | — | 97.0% |
Host prices and uptime from OpenRouter, synced 2026-10-22. Lower-precision weights (fp4, fp8) can cost less and score worse than the same model at bf16 — the score above is not host-specific.
Compare gpt-oss-20b
FAQ
How much does gpt-oss-20b cost?
gpt-oss-20b costs $0.018 per million input tokens and $0.090 per million output tokens on OpenRouter. At a 3:1 input:output mix that blends to $0.036 per million tokens.
What is gpt-oss-20b's context window?
gpt-oss-20b accepts up to 131,072 tokens of context, and can generate up to 32,768 output tokens.
How good is gpt-oss-20b on benchmarks?
gpt-oss-20b scores 60.8% on GPQA Diamond, independently run and published by Epoch AI. 2 benchmark results are listed on this page, each with its run date and source.