Model comparison
Claude Opus 4.6 vs Devstral 2
Claude Opus 4.6 has enough public results to be ranked (#20); Devstral 2 does not yet, so treat this comparison as directional.
Last verified . 2 shared benchmarks.
Mistral AISummary
- They share 2 benchmarks with published results for both. Claude Opus 4.6 scores higher in 1 category and Devstral 2 in 0 categories; one gap is clear of the uncertainty.
- The widest gap is in coding, where Claude Opus 4.6 leads 57.2 to 33.2.
- The biggest single-benchmark swing is SWE-bench Verified (bash only): 75.6% for Claude Opus 4.6 and 53.8% for Devstral 2.
- Devstral 2 is cheaper at $0.40 / $2 per million input/output tokens, against $5 / $25 for Claude Opus 4.6.
- Claude Opus 4.6 accepts more context: 1M tokens versus 262K.
- Devstral 2 has downloadable open weights; the other is API-only.
Side by side
| Claude Opus 4.6 | Devstral 2 | |
|---|---|---|
| Provider | Anthropic | Mistral AI |
| Noometry Index | 58.2 | 36.1 |
| Released | 2026-02-04 | 2025-12-09 |
| Weights | Proprietary | Open |
| Context window | 1M | 262K |
| Max output | 128K | 262K |
| Input $ / M tokens | $5 | $0.40 |
| Output $ / M tokens | $25 | $2 |
| Results tracked | 68 | 2 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Opus 4.6 leads
Claude Opus 4.6: 57.2 (#20), Devstral 2: 33.2 (#261)
| Benchmark | Claude Opus 4.6 | Devstral 2 |
|---|---|---|
| SWE-bench Verified (bash only) | 75.6% | 53.8% |
| LMArena WebDev | 1547 | 1197 |
| SWE-bench Verified | 78.7% | — |
| FrontierCode | 26.6% | — |
| SWE-bench Multilingual | 72% | — |
| GSO | 41.2% | — |
| WeirdML | 78% | — |
| LMArena Coding | 1536 | — |
| ALE-Bench | 996.5 | — |
| AlgoTune | 1.47 | — |
Agentic & Tool Use Not comparable
Claude Opus 4.6: 51.1 (#4), Devstral 2: —
| Benchmark | Claude Opus 4.6 | Devstral 2 |
|---|---|---|
| Terminal-Bench | 79.8% | — |
| APEX-Agents | 46.3% | — |
| Remote Labor Index | 4.2% | — |
| τ²-bench Banking | 27.3% | — |
| Cybench | 93% | — |
| DeepResearch Bench | 55.3% | — |
| GBAEval | 44.1% | — |
| LMArena Search | 1253 | — |
| METR Time Horizons | 78.9% | — |
| Vending-Bench 2 | 8,018 | — |
Reasoning Not comparable
Claude Opus 4.6: 57.8 (#23), Devstral 2: —
| Benchmark | Claude Opus 4.6 | Devstral 2 |
|---|---|---|
| ARC-AGI-2 | 69.2% | — |
| SimpleBench | 67.6% | — |
| Kagi LLM Benchmark | 83.6% | — |
| NYT Connections (extended) | 92.1% | — |
| ARC-AGI-1 | 94% | — |
| Chess Puzzles | 17% | — |
| EnigmaEval | 7.6% | — |
| Thematic Generalization | 80.6% | — |
| EBR-Bench | 12.7% | — |
| LMArena Hard Prompts | 1527 | — |
| Mystery Game Puzzles | 25% | — |
| DTBench | 91.2% | — |
| LMCA | 55.8% | — |
| Epoch Capabilities Index | 155.24 | — |
| ForecastBench | 60 | — |
Math Not comparable
Claude Opus 4.6: 63.0 (#31), Devstral 2: —
| Benchmark | Claude Opus 4.6 | Devstral 2 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 66% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 78.5% | — |
| OTIS Mock AIME 2024-2025 | 94.4% | — |
| ProofBench | 50% | — |
| LMArena Math | 1519 | — |
| FrontierMath (Feb 2025 set) | 40.7% | — |
| FrontierMath Tier 4 (v1) | 22.9% | — |
Knowledge Not comparable
Claude Opus 4.6: 61.9 (#26), Devstral 2: —
| Benchmark | Claude Opus 4.6 | Devstral 2 |
|---|---|---|
| GPQA Diamond | 90.5% | — |
| Humanity's Last Exam | 34.4% | — |
| SimpleQA Verified | 47% | — |
| Vectara Hallucination Rate | 12.2% | — |
| LMArena Expert | 1546 | — |
Multimodal Not comparable
Claude Opus 4.6: 37.3 (#74), Devstral 2: —
| Benchmark | Claude Opus 4.6 | Devstral 2 |
|---|---|---|
| LMArena Vision | 1316 | — |
| Furniture Assembly | 28.3% | — |
| LMArena Document | 1507 | — |
Multilingual Not comparable
Claude Opus 4.6: 57.9 (#6), Devstral 2: —
| Benchmark | Claude Opus 4.6 | Devstral 2 |
|---|---|---|
| LMArena Non-English | 1489 | — |
| LMArena Chinese | 1551 | — |
| LMArena French | 1513 | — |
| LMArena German | 1502 | — |
| LMArena Japanese | 1484 | — |
| LMArena Korean | 1464 | — |
| LMArena Russian | 1497 | — |
| LMArena Spanish | 1510 | — |
Instruction Following Not comparable
Claude Opus 4.6: 79.5 (#4), Devstral 2: —
| Benchmark | Claude Opus 4.6 | Devstral 2 |
|---|---|---|
| LMArena Instruction Following | 1523 | — |
Long Context Not comparable
Claude Opus 4.6: 48.1 (#13), Devstral 2: —
| Benchmark | Claude Opus 4.6 | Devstral 2 |
|---|---|---|
| CL-bench | 20.7% | — |
| CL-bench Life | 17% | — |
| LMArena Longer Query | 1520 | — |
Writing & Preference Not comparable
Claude Opus 4.6: 73.5 (#10), Devstral 2: —
| Benchmark | Claude Opus 4.6 | Devstral 2 |
|---|---|---|
| LMArena Text | 1503 | — |
| LMArena Creative Writing | 1505 | — |
| EQ-Bench Creative Writing | 1809 | — |
| EQ-Bench 4 | 1223 | — |
| LMArena Multi-Turn | 1513 | — |
Frequently asked questions
Is Claude Opus 4.6 better than Devstral 2?
Claude Opus 4.6 has enough public results to be ranked (#20); Devstral 2 does not yet, so treat this comparison as directional.
Which is cheaper, Claude Opus 4.6 or Devstral 2?
Devstral 2 is cheaper. It lists at $0.40 per million input tokens and $2 per million output tokens; Claude Opus 4.6 lists at $5 and $25.
Is Claude Opus 4.6 or Devstral 2 better for coding?
Claude Opus 4.6 scores higher on coding benchmarks: 57.2 versus 33.2 in the Noometry coding category.
Which has the bigger context window?
Claude Opus 4.6 does, with 1M tokens against 262K.
How many benchmarks do Claude Opus 4.6 and Devstral 2 share?
2 benchmarks have published results for both models. Claude Opus 4.6 has 68 scored results on Noometry and Devstral 2 has 2.