Model comparison
Codestral vs DeepSeek-V3.1-Terminus
DeepSeek-V3.1-Terminus is the stronger model overall, scoring 43.1 to 30.6 on the Noometry Index.
Last verified . 2 shared benchmarks.
Mistral AI
DeepSeekSummary
- They share 2 benchmarks with published results for both. Codestral scores higher in 0 categories and DeepSeek-V3.1-Terminus in 2 categories; 2 gaps are clear of the uncertainty.
- The widest gap is in coding, where DeepSeek-V3.1-Terminus leads 42.0 to 27.3.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 32.5% for Codestral and 57.4% for DeepSeek-V3.1-Terminus.
- Both cost about the same: $0.30 input and $0.90 output per million tokens.
- Codestral accepts more context: 256K tokens versus 164K.
- DeepSeek-V3.1-Terminus has downloadable open weights; the other is API-only.
Side by side
| Codestral | DeepSeek-V3.1-Terminus | |
|---|---|---|
| Provider | Mistral AI | DeepSeek |
| Noometry Index | 30.6 | 43.1 |
| Released | 2024-05-29 | 2025-09-22 |
| Weights | Proprietary | Open |
| Context window | 256K | 164K |
| Max output | 8K | 147K |
| Input $ / M tokens | $0.30 | $0.27 |
| Output $ / M tokens | $0.90 | $1 |
| Results tracked | 7 | 16 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.1-Terminus leads
Codestral: 27.3 (#321), DeepSeek-V3.1-Terminus: 42.0 (#113)
| Benchmark | Codestral | DeepSeek-V3.1-Terminus |
|---|---|---|
| ALE-Bench | 137.78 | 745.17 |
| Aider Polyglot | 11.1% | — |
| SciCode | — | 40.6% |
| BigCodeBench Instruct | 41.8% | — |
| LMArena Coding | — | 1426 |
| BigCodeBench Complete | 52.5% | — |
| HumanEval+ | 73.8% | — |
| MBPP+ | 61.9% | — |
Reasoning DeepSeek-V3.1-Terminus leads
Codestral: 19.8 (#251), DeepSeek-V3.1-Terminus: 26.4 (#133)
| Benchmark | Codestral | DeepSeek-V3.1-Terminus |
|---|---|---|
| Kagi LLM Benchmark | 32.5% | 57.4% |
| CritPt | — | 1.7% |
| LMArena Hard Prompts | — | 1426 |
| DTBench | — | 81.3% |
| LMCA | — | 28.6% |
Math Not comparable
Codestral: —, DeepSeek-V3.1-Terminus: 38.5 (#137)
| Benchmark | Codestral | DeepSeek-V3.1-Terminus |
|---|---|---|
| LMArena Math | — | 1402 |
Multilingual Not comparable
Codestral: —, DeepSeek-V3.1-Terminus: 52.1 (#92)
| Benchmark | Codestral | DeepSeek-V3.1-Terminus |
|---|---|---|
| LMArena Non-English | — | 1407 |
| LMArena Russian | — | 1436 |
Instruction Following Not comparable
Codestral: —, DeepSeek-V3.1-Terminus: 74.0 (#106)
| Benchmark | Codestral | DeepSeek-V3.1-Terminus |
|---|---|---|
| LMArena Instruction Following | — | 1404 |
Long Context Not comparable
Codestral: —, DeepSeek-V3.1-Terminus: 43.4 (#97)
| Benchmark | Codestral | DeepSeek-V3.1-Terminus |
|---|---|---|
| LMArena Longer Query | — | 1421 |
Writing & Preference Not comparable
Codestral: —, DeepSeek-V3.1-Terminus: 61.0 (#92)
| Benchmark | Codestral | DeepSeek-V3.1-Terminus |
|---|---|---|
| LMArena Text | — | 1419 |
| LMArena Creative Writing | — | 1403 |
| LMArena Multi-Turn | — | 1411 |
Frequently asked questions
Is Codestral better than DeepSeek-V3.1-Terminus?
DeepSeek-V3.1-Terminus is the stronger model overall, scoring 43.1 to 30.6 on the Noometry Index.
Which is cheaper, Codestral or DeepSeek-V3.1-Terminus?
Codestral is cheaper. It lists at $0.30 per million input tokens and $0.90 per million output tokens; DeepSeek-V3.1-Terminus lists at $0.27 and $1.
Is Codestral or DeepSeek-V3.1-Terminus better for coding?
DeepSeek-V3.1-Terminus scores higher on coding benchmarks: 42.0 versus 27.3 in the Noometry coding category.
Which has the bigger context window?
Codestral does, with 256K tokens against 164K.
How many benchmarks do Codestral and DeepSeek-V3.1-Terminus share?
2 benchmarks have published results for both models. Codestral has 7 scored results on Noometry and DeepSeek-V3.1-Terminus has 16.