GPT-5.1-Codex
OpenAI · released Nov 13, 2025
GPT-5.1-Codex is a specialized version of GPT-5.1 optimized for software engineering and coding workflows. It is designed for both interactive development sessions and long, independent execution of complex engineering tasks....
Specification
- Context window
- 400K
- Max output
- 128K
- Knowledge cutoff
- Not stated
- Parameters
- Undisclosed
- Licence
- Proprietary
- Serving providers
- 1
- Moderated
- No
Intelligence
35.6
82th percentile
Coding
—
Coding Index
Agentic
—
Agentic Index
Output speed
31 t/s
Median across providers
Latency
3.71s
Time to first token
Cost per task
$0.86
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 202595.7%
- GPQA Diamond86.0%
- MMLU-Pro86.0%
- LiveCodeBench84.9%
- τ²-bench (Telecom)83.0%
- IFBench70.0%
- AA-LCR (long context)69.0%
- SciCode40.2%
- Terminal-Bench Hard34.8%
- Humanity's Last Exam25.7%
An evaluation missing from this list was not run for this model — it is not a zero.
View as table
| Evaluation | Score |
|---|---|
| AIME 2025 | 95.7% |
| GPQA Diamond | 86.0% |
| MMLU-Pro | 86.0% |
| LiveCodeBench | 84.9% |
| τ²-bench (Telecom) | 83.0% |
| IFBench | 70.0% |
| AA-LCR (long context) | 69.0% |
| SciCode | 40.2% |
| Terminal-Bench Hard | 34.8% |
| Humanity's Last Exam | 25.7% |
Against its peers
Intelligence Index · this model highlighted, nearest peers in grey
- Kimi K2.5Moonshot AI36.0
- GPT-5.1-CodexOpenAI35.6
- Claude Opus 4.5Anthropic35.6
- GLM 5V TurboZ.ai35.3
- GPT-5OpenAI35.3
- Qwen3.5-27BQwen34.6
- M2.5MiniMax34.5
- GLM 4.7Z.ai34.5
Peers are the models sitting closest on the Intelligence Index — the set you would realistically choose between.
View as table
| Model | Intelligence |
|---|---|
| Kimi K2.5 | 36.0 |
| GPT-5.1-Codex | 35.6 |
| Claude Opus 4.5 | 35.6 |
| GLM 5V Turbo | 35.3 |
| GPT-5 | 35.3 |
| Qwen3.5-27B | 34.6 |
| M2.5 | 34.5 |
| GLM 4.7 | 34.5 |
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
- $1.25
- Output / 1M tokens
- $10
- Cached input / 1M
- $0.13
- Blended 3:1
- $3.44
One task, estimated
$0.86
- 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
- 1,205
- Win rate
- 51.4%
- Strongest at
- Data visualisation
- Tournaments
- 70
Elo by category
Dot position on a 1,175–1,225 scale · Elo has no meaningful zero
- 1,205
- 1,183
- 1,180
Agent categories (full-stack apps, mobile apps) are only scored for models tested in agent harnesses.
View as table
| Category | Elo | Win rate |
|---|---|---|
| Data visualisation | 1205 | 51.4% |
| Websites | 1183 | 56.0% |
| Game development | 1180 | 52.3% |