Model Details
GPT-5.1 Codex
Quick answer: GPT-5.1 Codex is OpenAI's February 2026 coding-specialised model, scoring 73.7% on SWE-Bench Verified. It is the updated successor to GPT-5 Codex with a more recent September 2025 knowledge cutoff.
At a Glance
Where GPT-5.1 Codex leads
- 73.7% SWE-Bench Verified — strong software engineering
- More recent knowledge cutoff (September 2025) vs GPT-5 Codex (September 2024)
- Latest Codex generation from OpenAI
Where it lags
- Slightly below GPT-5 Codex on SWE-Bench (73.7% vs 74.5%)
- Below DeepSeek-V4 Flash Max (79%) on SWE-Bench
- Proprietary: no open weights
Best for: Current-generation coding tasks via the OpenAI API; SWE-Bench class software engineering; automated coding agents.
What GPT-5.1 Codex Is
GPT-5.1 Codex (released February 2026) is OpenAI's updated Codex model — the February 2026 iteration of the coding-specialised GPT-5 Codex family. The ".1" suffix indicates an incremental update, with a significantly more recent knowledge cutoff (September 2025 vs September 2024).
The model's 73.7% SWE-Bench Verified score is marginally below GPT-5 Codex (74.5%), suggesting evaluation methodology or benchmark drift rather than capability regression. Both models are competitive in the 73-75% SWE-Bench range.
Specifications
| Field | Value |
|---|---|
| Organization | OpenAI |
| License | Proprietary (API only) |
| Release date | February 2026 |
| Knowledge cutoff | September 2025 |
| Modality | Text / Code |
Pricing
Available via OpenAI API — refer to the OpenAI pricing page.
Public Benchmark Scores
| Benchmark | Score | Source | Date |
|---|---|---|---|
| SWE-Bench Verified | 73.7% | Benchgen evaluation | 2026-02 |
GPT-5.1 Codex vs Alternatives
| Model | SWE-Bench Verified | Knowledge Cutoff |
|---|---|---|
| GPT-5.1 Codex | 73.7% | Sep 2025 |
| GPT-5 Codex | 74.5% | Sep 2024 |
| DeepSeek-V4 Flash Max | 79% | Sep 2025 |
GPT-5.1 Codex is preferred over GPT-5 Codex for current-date code knowledge (Sep 2025 cutoff). For maximum SWE-Bench: DeepSeek-V4 Flash Max (79%, proprietary).
Related Models & Benchmarks
- Compare: GPT-5 Codex · DeepSeek-V3.2 Speciale · DeepSeek-V4 Flash Max
- Benchmarks: SWE-Bench Verified
Frequently Asked Questions
What is GPT-5.1 Codex?
OpenAI's February 2026 updated coding model scoring 73.7% SWE-Bench Verified with a September 2025 knowledge cutoff.Should I use GPT-5 Codex or GPT-5.1 Codex?
GPT-5.1 Codex for most uses — more recent knowledge cutoff (Sep 2025 vs Sep 2024). GPT-5 Codex has marginally higher SWE-Bench (74.5% vs 73.7%).Specs from OpenAI's GPT-5.1 Codex release (February 2026) and Benchgen evaluations. Last updated 2026-07-24.
Benchmark Leaderboards
This model isn’t on any benchmark leaderboard yet.