Self-hosting and data control
Choose gpt-oss-120b
gpt-oss-120b has open weights, so you can run it on your own servers; Codestral is only available through its provider.
Codestral vs gpt-oss-120b: price, context and capabilities
Choose gpt-oss-120b for self-hosting and data control; on the other points the two are level or the data is missing.
In short
- A price comparison is not possible: one of them has no comparable published price.
- gpt-oss-120b reads more per request: 131k tokens against 128k.
- Only gpt-oss-120b offers reasoning, self-hostable.
Which should you choose?
Each answer below comes straight from the published prices, specifications and independent benchmark results.
Specifications side by side
| Codestral | gpt-oss-120b | |
|---|---|---|
| Provider | Mistral AI | OpenAI |
| Category | Language models | Language models |
| Status | Generally available | Generally available |
| Headline price | No data | No data |
| Input | No data | No data |
| Output | No data | No data |
| Cached input | No data | No data |
| Context window | 128k | 131k |
| Max output | No data | 131k |
| Released | August 2025 | August 2025 |
| Knowledge cutoff | No data | June 2024 |
| Licence | Proprietary | Open weights |
| Capabilities | Tool calling | Reasoning, Tool calling, Self-hostable |
| API model id | codestral-2508 | gpt-oss-120b |
Benchmarks side by side
Leader in bold.
| Benchmark | Codestral | gpt-oss-120b |
|---|---|---|
| Intelligence Index | No data | 11.6 |
| Terminal-Bench Hard | No data | 23.5% |
| SciCode | No data | 34% |
| GPQA Diamond | No data | 78.2% |
| Humanity's Last Exam | No data | 19.6% |
| IFBench (instruction following) | No data | 69% |
| τ²-Bench (agentic tool use) | No data | 65.8% |
| Long-context reasoning (AA-LCR) | No data | 52% |
Independent evaluations by Artificial Analysis, run the same way for every model, at the reasoning setting shown. Percentages unless marked as an index.
Frequently asked questions
Should I choose Codestral or gpt-oss-120b?
Choose gpt-oss-120b for self-hosting and data control; on the other points the two are level or the data is missing.
Which is cheaper, Codestral or gpt-oss-120b?
A price comparison is not possible: one of them has no comparable published price.
Which has the larger context window, Codestral or gpt-oss-120b?
gpt-oss-120b reads more per request: 131k tokens against 128k.
What is the difference between Codestral and gpt-oss-120b?
Codestral is made by Mistral AI and gpt-oss-120b by OpenAI. A price comparison is not possible: one of them has no comparable published price. gpt-oss-120b reads more per request: 131k tokens against 128k. Only gpt-oss-120b offers reasoning, self-hostable.