Compare / head-to-head
Devstral Small 2vs
MiniMax M2
MiniMax M2 leads 2 of 2 shared benchmarks. Devstral Small 2 is 3.8x cheaper per token. Devstral Small 2 has the larger context window (256K tokens).
Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.
Shared benchmarks
0 – 2
MiniMax M2 leads
Cheaper per token
Devstral Small 2
3.8x cheaper, input + output
Larger context
Devstral Small 2
256K tokens
Providers
2 providers
Mistral AI · MiniMax
Devstral Small 2
Mistral AI · released 2025-12-09
textvision
MiniMax M2
MiniMax · released 2025-10-27
text
Quality
Benchmark matrix
2 shared · 1 only Devstral Small 2 · 15 only MiniMax M2| Benchmark | Devstral Small 2 | MiniMax M2 | Δ | Edge |
|---|---|---|---|---|
| Reported by both · 2 | ||||
| SWE-bench Verified | 68% ↗ | 69.4% ↗ | -1.4 pt | MiniMax M2 |
| SWE-bench Multilingual | 55.7% ↗ | 56.5% ↗ | -0.8 pt | MiniMax M2 |
| Only Devstral Small 2 reports · 1 | ||||
| Terminal Bench 2 | 22.5% ↗ | not reported | — | — |
| Only MiniMax M2 reports · 15 | ||||
| LMArena Elo | not reported | 1343.4 ↗ | — | — |
| AgentCompany | not reported | 36% ↗ | — | — |
| ArtifactsBench | not reported | 66.8% ↗ | — | — |
| BrowseComp | not reported | 44% ↗ | — | — |
| BrowseComp-zh | not reported | 48.5% ↗ | — | — |
| FinSearchComp-global | not reported | 65.5% ↗ | — | — |
| GAIA (text only) | not reported | 75.7% ↗ | — | — |
| HLE (w/ tools) | not reported | 31.8% ↗ | — | — |
| LMArena WebDev (LMArena) | not reported | 1296.6 ↗ | — | — |
| Multi-SWE-Bench | not reported | 36.2% ↗ | — | — |
| SimpleBench (SimpleBench) | not reported | 25% ↗ | — | — |
| Terminal-Bench | not reported | 46.3% ↗ | — | — |
| Vending-Bench 2 (Andon Labs) | not reported | 160.6 ↗ | — | — |
| xbench-DeepSearch | not reported | 72% ↗ | — | — |
| τ²-Bench | not reported | 77.2% ↗ | — | — |
Scores tagged "epoch" or "matharena" are independent runs, used only where the lab has not published its own; ⚠ marks rows MathArena flags as released after the competition. Higher is better on every row. Δ is Devstral Small 2 minus MiniMax M2 in the benchmark's own unit. "Not reported" means the lab has not published that figure; it is not a zero. ↗ opens the source.
Specs & pricing
Side by side
Official pricing pages and model cards| Spec | Devstral Small 2 | MiniMax M2 | Edge |
|---|---|---|---|
| Context window | 256K tokens | 205K tokens | Devstral Small 2 |
| Max output | — | 128K tokens | — |
| Input price / 1M | $0.1 ↗ | $0.3 ↗ | Devstral Small 2 |
| Output price / 1M | $0.3 ↗ | $1.2 ↗ | Devstral Small 2 |
| Cached input / 1M | — | $0.03 ↗ | — |
| Input + output / 1M Lower is cheaper. List prices; batch, tool and regional fees excluded. | $0.4 | $1.5 | Devstral Small 2 |
| Modalities | text · vision | text | Devstral Small 2 |
| Released | 2025-12-09 | 2025-10-27 | — |
| Cited benchmark scores | 3 | 17 | — |
Reliability
Provider status
Live from /statusMore matchups
Devstral Small 2 vs …
Models sharing the most benchmarksMore matchups
MiniMax M2 vs …
Models sharing the most benchmarksBuilt by Respan
Which one wins on your data?
Public benchmarks are a starting point. Run Devstral Small 2 and MiniMax M2 on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.