Model comparison
Codestral vs MiniMax-M2.5
MiniMax-M2.5 is the stronger model overall, scoring 38.3 to 30.6 on the Noometry Index.
Last verified . 2 shared benchmarks.
Mistral AI
MiniMaxSummary
- They share 2 benchmarks with published results for both. Codestral scores higher in 1 category and MiniMax-M2.5 in 1 category; 2 gaps are clear of the uncertainty.
- The widest gap is in coding, where MiniMax-M2.5 leads 48.1 to 27.3.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 32.5% for Codestral and 55.2% for MiniMax-M2.5.
- Codestral is cheaper at $0.30 / $0.90 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M2.5.
- Codestral accepts more context: 256K tokens versus 205K.
- MiniMax-M2.5 has downloadable open weights; the other is API-only.
Side by side
| Codestral | MiniMax-M2.5 | |
|---|---|---|
| Provider | Mistral AI | MiniMax |
| Noometry Index | 30.6 | 38.3 |
| Released | 2024-05-29 | 2026-02-12 |
| Weights | Proprietary | Open |
| Context window | 256K | 205K |
| Max output | 8K | 131K |
| Input $ / M tokens | $0.30 | $0.30 |
| Output $ / M tokens | $0.90 | $1.20 |
| Results tracked | 7 | 33 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M2.5 leads
Codestral: 27.3 (#321), MiniMax-M2.5: 48.1 (#58)
| Benchmark | Codestral | MiniMax-M2.5 |
|---|---|---|
| ALE-Bench | 137.78 | 618.17 |
| SWE-bench Verified (bash only) | — | 75.8% |
| Aider Polyglot | 11.1% | — |
| LMArena WebDev | — | 1387 |
| SWE-bench Multilingual | — | 68.3% |
| BigCodeBench Instruct | 41.8% | — |
| LMArena Coding | — | 1381 |
| BigCodeBench Complete | 52.5% | — |
| HumanEval+ | 73.8% | — |
| MBPP+ | 61.9% | — |
Agentic & Tool Use Not comparable
Codestral: —, MiniMax-M2.5: 30.4 (#77)
| Benchmark | Codestral | MiniMax-M2.5 |
|---|---|---|
| Terminal-Bench | — | 42.7% |
| Vending-Bench 2 | — | -23.16 |
Reasoning Codestral leads
Codestral: 19.8 (#251), MiniMax-M2.5: 17.5 (#292)
| Benchmark | Codestral | MiniMax-M2.5 |
|---|---|---|
| Kagi LLM Benchmark | 32.5% | 55.2% |
| ARC-AGI-2 | — | 4.9% |
| NYT Connections (extended) | — | 16.8% |
| ARC-AGI-1 | — | 63.7% |
| LMArena Hard Prompts | — | 1372 |
| Epoch Capabilities Index | — | 146.68 |
Math Not comparable
Codestral: —, MiniMax-M2.5: 26.9 (#253)
| Benchmark | Codestral | MiniMax-M2.5 |
|---|---|---|
| ProofBench | — | 4% |
| LMArena Math | — | 1378 |
Knowledge Not comparable
Codestral: —, MiniMax-M2.5: 39.2 (#135)
| Benchmark | Codestral | MiniMax-M2.5 |
|---|---|---|
| Vectara Hallucination Rate | — | 9.1% |
| LMArena Expert | — | 1379 |
Multilingual Not comparable
Codestral: —, MiniMax-M2.5: 47.1 (#152)
| Benchmark | Codestral | MiniMax-M2.5 |
|---|---|---|
| LMArena Non-English | — | 1338 |
| LMArena Chinese | — | 1393 |
| LMArena French | — | 1362 |
| LMArena German | — | 1362 |
| LMArena Japanese | — | 1171 |
| LMArena Korean | — | 1232 |
| LMArena Russian | — | 1358 |
| LMArena Spanish | — | 1354 |
Instruction Following Not comparable
Codestral: —, MiniMax-M2.5: 71.5 (#148)
| Benchmark | Codestral | MiniMax-M2.5 |
|---|---|---|
| LMArena Instruction Following | — | 1353 |
Long Context Not comparable
Codestral: —, MiniMax-M2.5: 37.5 (#216)
| Benchmark | Codestral | MiniMax-M2.5 |
|---|---|---|
| CL-bench | — | 11.4% |
| CL-bench Life | — | 6.3% |
| LMArena Longer Query | — | 1366 |
Writing & Preference Not comparable
Codestral: —, MiniMax-M2.5: 53.9 (#153)
| Benchmark | Codestral | MiniMax-M2.5 |
|---|---|---|
| LMArena Text | — | 1359 |
| LMArena Creative Writing | — | 1331 |
| EQ-Bench Creative Writing | — | 1361 |
| LMArena Multi-Turn | — | 1364 |
Frequently asked questions
Is Codestral better than MiniMax-M2.5?
MiniMax-M2.5 is the stronger model overall, scoring 38.3 to 30.6 on the Noometry Index.
Which is cheaper, Codestral or MiniMax-M2.5?
Codestral is cheaper. It lists at $0.30 per million input tokens and $0.90 per million output tokens; MiniMax-M2.5 lists at $0.30 and $1.20.
Is Codestral or MiniMax-M2.5 better for coding?
MiniMax-M2.5 scores higher on coding benchmarks: 48.1 versus 27.3 in the Noometry coding category.
Which has the bigger context window?
Codestral does, with 256K tokens against 205K.
How many benchmarks do Codestral and MiniMax-M2.5 share?
2 benchmarks have published results for both models. Codestral has 7 scored results on Noometry and MiniMax-M2.5 has 33.