Building more with GPT-5.1-Codex-Max (OpenAI) Nov 19, 2025
GPT-5.1-Codex-Max
GPT-5.1-Codex-Max ranks first on 2 of the 25 benchmarks with five or more models scored.
28
Benchmarks with a score
2
Ranked first, out of five or more models
6
In the top three
OpenAI Nov 2025
GPT-5.1-Codex-Max
Closed weights
150.5
#88 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
- Terminal-Bench 2.0 60.5% #12 /43
- PostTrainBench 19.7% #10 /23
- SWE-Lancer 80% #2 /10
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ProofBench | 9.00% | #46 of 59 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| METR Time Horizon | 2.7 h | #2 of 31 | Official leaderboard | |
| OpenAI-Proof Q&A | 8.00% | #1 of 9 | Lab model card | |
| SWE-bench Verified | 77.9% | #29 of 155 | Lab model card | |
| SWE-Lancer | 80% | #2 of 10 | Lab model card | |
| OpenAI PRs | 53% | #3 of 12 | Lab model card | |
| METR Time Horizon 1.1 | 3.7 h | #9 of 26 | Official leaderboard | |
| Vibe Code Bench v1 | 22.2% | #6 of 17 | Official leaderboard | |
| MLE-bench 30 | 17% | #3 of 8 | Lab model card | |
| IOI (Vals) | 21.4% | #22 of 55 | Official leaderboard | |
| ALE-Bench | 1,538 | #53 of 126 | Official leaderboard | |
| PostTrainBench | 19.7% | #10 of 23 | Official leaderboard | |
| PaperBench | 40% | #7 of 15 | Lab model card | |
| OTelBench | 11.6% | #9 of 14 | Official leaderboard | |
| Vibe Code Bench | 22.2% | #85 of 120 | Official leaderboard |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SnakeBench | 38 | #1 of 188 | Official leaderboard | |
| Terminal-Bench 2.0 | 60.5% | #12 of 43 | Official leaderboard |
Safety and security
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Coding prompt injections | 100% | #1 of 4 | Lab model card | |
| Tacit Knowledge and Troubleshooting | 77% | #10 of 30 | Lab model card | |
| Capture the Flag (Professional) | 76% | #8 of 21 | Lab model card | |
| StrongREJECT | 96.7% | #8 of 19 | Lab model card | |
| ProtocolQA Open-Ended | 35% | #15 of 27 | Lab model card | |
| Malware Refusals (Golden Set) | 100% | #2 of 3 | Lab model card | |
| CVE-Bench | 80% | #10 of 14 | Lab model card | |
| TroubleshootingBench | 29% | #16 of 20 | Lab model card | |
| Destructive action avoidance | 75% | #7 of 8 | Lab model card | |
| Multimodal Troubleshooting Virology | 38% | #27 of 27 | Lab model card | |
| Long-form Biological Risk Questions | 0% | #3 of 3 | Lab model card |
Its lab's numbers put it at 150.5, number 88 of 404
The Self-Reported Index fits one score to the 7 capability results OpenAI reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 146.8 to 154.2. How we compute it.
- SWE-Lancer 80
- SWE-bench Verified 77.9
- Terminal-Bench 2.0 58.1
- OpenAI PRs 53
- PaperBench 40
- MLE-bench 30 17
- OpenAI-Proof Q&A 8
Its lab's number next to a leaderboard
Where the benchmark's own leaderboard, or Epoch AI's run, also has this model. Gaps mostly come from different harnesses, effort levels and task subsets. Every pair.
| Benchmark | Lab says | Board says | Gap |
|---|---|---|---|
| Terminal-Bench 2.0 | 58.1% Terminal-Bench 2.0, xhigh | 60.5% Leaderboard, Codex CLI | −2.4 pts |
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.
Its system card
Every result in the longer documents the lab published with it. 20 of them aren't in the launch post.
GPT-5.1-Codex-Max System Card Nov 18, 2025
Safety
Robustness
- StrongREJECT 96.7
- Coding prompt injections 100
Alignment
Bio and chem risk
Cyber risk
AI R&D risk
- SWE-Lancer 80
- PaperBench 40
- MLE-bench 30 17
- OpenAI PRs 53
- OpenAI-Proof Q&A 8