gpt-oss-20b
OpenAI · released Aug 5, 2025
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...
Specification
- Context window
- 131K
- Max output
- 33K
- Knowledge cutoff
- Jun 30, 2024
- Parameters
- 20.9B
- Licence
- apache-2.0
- Serving providers
- 10
- Moderated
- No
- Uptime
- 100.0%
Intelligence
9.0
37th percentile
Coding
20.7
Coding Index
Agentic
1.2
Agentic Index
Output speed
410 t/s
Median across providers
Latency
225ms
Time to first token
Cost per task
$0.008
Estimated
Benchmarks
Where the score comes from
The Intelligence Index is a composite. These are the underlying evaluations this model was actually measured on.
Evaluation scores
Percentage correct · higher is better
- AIME 202589.3%
- LiveCodeBench77.7%
- MMLU-Pro74.8%
- GPQA Diamond68.8%
- IFBench65.1%
- τ²-bench (Telecom)60.2%
- SciCode38.9%
- AA-LCR (long context)34.7%
- Humanity's Last Exam11.0%
- Terminal-Bench Hard10.6%
An evaluation missing from this list was not run for this model — it is not a zero.
View as table
| Evaluation | Score |
|---|---|
| AIME 2025 | 89.3% |
| LiveCodeBench | 77.7% |
| MMLU-Pro | 74.8% |
| GPQA Diamond | 68.8% |
| IFBench | 65.1% |
| τ²-bench (Telecom) | 60.2% |
| SciCode | 38.9% |
| AA-LCR (long context) | 34.7% |
| Humanity's Last Exam | 11.0% |
| Terminal-Bench Hard | 10.6% |
Against its peers
Intelligence Index · this model highlighted, nearest peers in grey
- 9.6
- Llama 4 MaverickMeta9.3
- Qwen3 Coder NextQwen9.2
- Mistral Medium 3.1Mistral AI9.2
- gpt-oss-20bOpenAI9.0
- Mistral Medium 3Mistral AI9.0
- LFM2.5-2.6B (free)Liquid8.4
- GLM 4.6VZ.ai8.4
Peers are the models sitting closest on the Intelligence Index — the set you would realistically choose between.
View as table
| Model | Intelligence |
|---|---|
| Qwen3 Next 80B A3B Instruct | 9.6 |
| Llama 4 Maverick | 9.3 |
| Qwen3 Coder Next | 9.2 |
| Mistral Medium 3.1 | 9.2 |
| gpt-oss-20b | 9.0 |
| Mistral Medium 3 | 9.0 |
| LFM2.5-2.6B (free) | 8.4 |
| GLM 4.6V | 8.4 |
Percentile among all indexed models
Pricing
What it costs to run
List prices per million tokens, plus what one representative task works out to.
List price
- Input / 1M tokens
- $0.018
- Output / 1M tokens
- $0.09
- Cached input / 1M
- $0.009
- Blended 3:1
- $0.036
One task, estimated
$0.008
- Input tokens
- 50,000
- Output tokens
- 80,000
- Profile
- Reasoning
Estimated from list pricing: 50K input tokens plus 80K output tokens for reasoning models (25K for non-reasoning).
Serving providers
Speed and latency figures are medians across these providers, so a widely-served model reports a blend rather than any single endpoint.
Arena
Head-to-head generation quality
Elo from pairwise judgements, broken out by the kind of thing the model was asked to build.
- Overall Elo
- 854
- Win rate
- 27.9%
- Strongest at
- Websites
- Tournaments
- 627
Elo by category
Dot position on a 850–875 scale · Elo has no meaningful zero
- 854
Agent categories (full-stack apps, mobile apps) are only scored for models tested in agent harnesses.
View as table
| Category | Elo | Win rate |
|---|---|---|
| Websites | 854 | 27.9% |