Mistral AIopen-weight
Released: 2025-01-14Codestral 25.01
Mistral dedicated 256k context window coding model tailored for fill-in-the-middle completion and polyglot development.
Intelligence Index
78.5/ 100
Calibrated multi-domain compositeThroughput Speed
90tok/sec
Output streaming throughputTime To First Token
320ms
Initial chunk server latencyLiveBench Contam-Free
—%
September 2026 suite scoreSourced Benchmark Evaluations (60)
Verified performance across authoritative benchmarks with provenance tracking.| Benchmark | Status | Score | Trust | Date | Source Type | Provenance |
|---|---|---|---|---|---|---|
| ARC-AGI-3 | Active | 48.1% | 93 | 2025-02-18 | independent | Source ↗ |
| ARC-AGI-2 | Nearing Saturation | 57.7% | 75 | 2025-02-11 | independent | Source ↗ |
| FrontierMath | Active | 19.9% | 98 | 2025-02-08 | independent | Source ↗ |
| PutnamBench | Active | 36.9% | 90 | 2025-02-07 | independent | Source ↗ |
| Omni-MATH | Deprecated | 52.5% | 10 | 2025-02-05 | independent | Source ↗ |
| SWE-Lancer | Deprecated | 49% | 20 | 2025-02-05 | independent | Source ↗ |
| FrontierCode | Active | 41.8% | 98 | 2025-02-03 | independent | Source ↗ |
| Soohak | Active | 54% | 98 | 2025-02-03 | independent | Source ↗ |
| StructFlowBench | Nearing Saturation | 61.4% | 75 | 2025-02-01 | independent | Source ↗ |
| Humanity's Last Exam | Active | 32.6% | 93 | 2025-02-01 | independent | Source ↗ |
| AIME | Nearing Saturation | 67.5% | 75 | 2025-02-01 | independent | Source ↗ |
| OlympiadBench | Active | 58.2% | 80 | 2025-01-31 | independent | Source ↗ |
| Terminal-Bench | Active | 50.4% | 80 | 2025-01-30 | independent | Source ↗ |
| FireBench | Active | 60.7% | 95 | 2025-01-30 | independent | Source ↗ |
| MathArena | Active | 56.8% | 99 | 2025-01-30 | independent | Source ↗ |
| SimpleQA | Active | 40.8% | 90 | 2025-01-29 | independent | Source ↗ |
| OpenAI-MRCR | Active | 65.9% | 80 | 2025-01-29 | independent | Source ↗ |
| CodeContests | Deprecated | 46.8% | 10 | 2025-01-29 | independent | Source ↗ |
| SWE-bench Pro | Active | 44.6% | 90 | 2025-01-28 | independent | Source ↗ |
| ComplexBench | Nearing Saturation | 59.2% | 80 | 2025-01-28 | independent | Source ↗ |
| LongBench-Pro | Active | 64.4% | 80 | 2025-01-28 | independent | Source ↗ |
| MultiPL-E | Deprecated | 66.2% | 10 | 2025-01-27 | independent | Source ↗ |
| ZeroSCROLLS | Nearing Saturation | 62.8% | 78 | 2025-01-27 | independent | Source ↗ |
| InfiniteBench | Nearing Saturation | 66.7% | 65 | 2025-01-26 | independent | Source ↗ |
| LiveCodeBench | Saturated | 59% | 39 | 2025-01-26 | independent | Source ↗ |
| AgentIF | Nearing Saturation | 49.1% | 80 | 2025-01-26 | independent | Source ↗ |
| MATH | Saturated | 69.6% | 25 | 2025-01-26 | independent | Source ↗ |
| SCROLLS | Nearing Saturation | 62.8% | 65 | 2025-01-25 | independent | Source ↗ |
| Mercury | Deprecated | 63.4% | 10 | 2025-01-25 | independent | Source ↗ |
| MMLU-Pro | Nearing Saturation | 72.2% | 75 | 2025-01-25 | independent | Source ↗ |
| TabMWP | Saturated | 65.3% | 25 | 2025-01-24 | independent | Source ↗ |
| CyberSecEval | Deprecated | 64.8% | 10 | 2025-01-24 | independent | Source ↗ |
| RULER | Nearing Saturation | 72.2% | 84 | 2025-01-24 | independent | Source ↗ |
| SWE-bench Verified | Saturated | 56.2% | 25 | 2025-01-23 | independent | Source ↗ |
| QuALITY | Nearing Saturation | 69.1% | 78 | 2025-01-23 | independent | Source ↗ |
| MMLU-Redux | Saturated | 76.9% | 25 | 2025-01-23 | independent | Source ↗ |
| SWE-bench | Deprecated | 49% | 10 | 2025-01-22 | independent | Source ↗ |
| MGSM | Nearing Saturation | 67.5% | 65 | 2025-01-22 | independent | Source ↗ |
| LongBench | Nearing Saturation | 67.5% | 65 | 2025-01-22 | independent | Source ↗ |
| QASPER | Nearing Saturation | 65.2% | 65 | 2025-01-21 | independent | Source ↗ |
| TruthfulQA | Saturated | 69.1% | 25 | 2025-01-21 | independent | Source ↗ |
| AGIEval | Saturated | 63.6% | 25 | 2025-01-21 | independent | Source ↗ |
| SuperGLUE | Saturated | 75.5% | 25 | 2025-01-20 | independent | Source ↗ |
| GSM-Hard | Deprecated | 68.2% | 10 | 2025-01-20 | independent | Source ↗ |
| MT-Bench | Nearing Saturation | 7.4% | 75 | 2025-01-19 | independent | Source ↗ |
| BIG-Bench Hard | Saturated | 72.5% | 25 | 2025-01-19 | independent | Source ↗ |
| CodeXGLUE | Deprecated | 67.7% | 23 | 2025-01-19 | independent | Source ↗ |
| WinoGrande | Saturated | 77% | 28 | 2025-01-18 | independent | Source ↗ |
| MBPP+ | Nearing Saturation | 73.4% | 65 | 2025-01-18 | independent | Source ↗ |
| IFEval | Deprecated | 70.8% | 25 | 2025-01-18 | independent | Source ↗ |
| GPQA | Nearing Saturation | 69.1% | 78 | 2025-01-18 | independent | Source ↗ |
| HumanEval+ | Saturated | 77.8% | 25 | 2025-01-17 | independent | Source ↗ |
| Needle-in-a-Haystack | Saturated | 95.7% | 44 | 2025-01-17 | independent | Source ↗ |
| CommonsenseQA | Saturated | 77.7% | 28 | 2025-01-17 | independent | Source ↗ |
| GSM8K | Saturated | 76.7% | 25 | 2025-01-16 | independent | Source ↗ |
| GLUE | Saturated | 79.9% | 28 | 2025-01-16 | independent | Source ↗ |
| MBPP | Saturated | 76.3% | 25 | 2025-01-16 | independent | Source ↗ |
| HumanEval | Saturated | 82.1% | 25 | 2025-01-15 | independent | Source ↗ |
| MMLU | Saturated | 80.1% | 25 | 2025-01-15 | independent | Source ↗ |
| ARC-AGI | Saturated | 69.6% | 25 | 2025-01-15 | independent | Source ↗ |
Local Hardware Execution (Ollama)Hardware Compatible
# 1. Pull and execute model locally with Ollama
ollama run codestral-25-01
# 2. Or invoke via local OpenAI-compatible endpoint
curl http://localhost:11434/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
"model": "codestral-25-01",
"messages": [{"role": "user", "content": "Analyze reasoning chains on AIME 2026"}]
}'Evaluation & Sourcing Notes
Scores listed for Codestral 25.01 represent verified evaluations extracted from official research papers, independent evaluation suites (HELM, LMSYS, OpenCompass, LiveBench), and verified audit reports.
All benchmarks marked as saturated or deprecated reflect historical performance where the benchmark no longer provides active discriminative power.