Model comparison
Devstral Small 2505 vs Step 3
Step 3 is the stronger model overall, scoring 39.9 to 33.9 on the Noometry Index.
Last verified . 1 shared benchmarks.
Mistral AI
StepFunSummary
- They share 1 benchmark with published results for both. Devstral Small 2505 scores higher in 0 categories and Step 3 in 2 categories; one gap is clear of the uncertainty.
- The widest gap is in reasoning, where Step 3 leads 27.6 to 19.1.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 37.7% for Devstral Small 2505 and 62.3% for Step 3.
Side by side
| Devstral Small 2505 | Step 3 | |
|---|---|---|
| Provider | Mistral AI | StepFun |
| Noometry Index | 33.9 | 39.9 |
| Released | 2025-05-07 | — |
| Weights | Open | Open |
| Context window | 128K | — |
| Max output | 128K | — |
| Input $ / M tokens | $0.10 | — |
| Output $ / M tokens | $0.30 | — |
| Results tracked | 4 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Devstral Small 2505: 38.6 (#160), Step 3: 39.5 (#144)
| Benchmark | Devstral Small 2505 | Step 3 |
|---|---|---|
| SWE-bench Verified (bash only) | 56.4% | — |
| SciCode | 28.8% | — |
| LMArena Coding | — | 1367 |
Reasoning Step 3 leads
Devstral Small 2505: 19.1 (#243), Step 3: 27.6 (#102)
| Benchmark | Devstral Small 2505 | Step 3 |
|---|---|---|
| Kagi LLM Benchmark | 37.7% | 62.3% |
| CritPt | 0% | — |
| LMArena Hard Prompts | — | 1355 |
Math Not comparable
Devstral Small 2505: —, Step 3: 34.9 (#148)
| Benchmark | Devstral Small 2505 | Step 3 |
|---|---|---|
| LMArena Math | — | 1366 |
Knowledge Not comparable
Devstral Small 2505: —, Step 3: 36.2 (#161)
| Benchmark | Devstral Small 2505 | Step 3 |
|---|---|---|
| LMArena Expert | — | 1333 |
Multimodal Not comparable
Devstral Small 2505: —, Step 3: 34.9 (#81)
| Benchmark | Devstral Small 2505 | Step 3 |
|---|---|---|
| LMArena Vision | — | 1176 |
Multilingual Not comparable
Devstral Small 2505: —, Step 3: 46.5 (#154)
| Benchmark | Devstral Small 2505 | Step 3 |
|---|---|---|
| LMArena Non-English | — | 1327 |
| LMArena Chinese | — | 1396 |
| LMArena German | — | 1371 |
| LMArena Korean | — | 1271 |
| LMArena Russian | — | 1331 |
| LMArena Spanish | — | 1372 |
Instruction Following Not comparable
Devstral Small 2505: —, Step 3: 70.6 (#159)
| Benchmark | Devstral Small 2505 | Step 3 |
|---|---|---|
| LMArena Instruction Following | — | 1332 |
Long Context Not comparable
Devstral Small 2505: —, Step 3: 40.3 (#152)
| Benchmark | Devstral Small 2505 | Step 3 |
|---|---|---|
| LMArena Longer Query | — | 1326 |
Writing & Preference Not comparable
Devstral Small 2505: —, Step 3: 54.5 (#146)
| Benchmark | Devstral Small 2505 | Step 3 |
|---|---|---|
| LMArena Text | — | 1351 |
| LMArena Creative Writing | — | 1322 |
| LMArena Multi-Turn | — | 1340 |
Frequently asked questions
Is Devstral Small 2505 better than Step 3?
Step 3 is the stronger model overall, scoring 39.9 to 33.9 on the Noometry Index.
Is Devstral Small 2505 or Step 3 better for coding?
They score almost the same on coding (38.6 vs 39.5); test both on your own repository before choosing.
How many benchmarks do Devstral Small 2505 and Step 3 share?
1 benchmark has published results for both models. Devstral Small 2505 has 4 scored results on Noometry and Step 3 has 17.