Model comparison
Devstral Small 2505 vs GPT-4.5
GPT-4.5 is the stronger model overall, scoring 37.2 to 34.3 on the Noometry Index.
Last verified . 0 shared benchmarks.
Mistral AISummary
- The widest gap is in reasoning, where Devstral Small 2505 leads 19.7 to 13.9.
- Devstral Small 2505 has downloadable open weights; the other is API-only.
Side by side
| Devstral Small 2505 | GPT-4.5 | |
|---|---|---|
| Provider | Mistral AI | OpenAI |
| Noometry Index | 34.3 | 37.2 |
| Released | 2025-05-07 | 2025-02-27 |
| Weights | Open | Proprietary |
| Context window | 128K | — |
| Max output | 128K | — |
| Input $ / M tokens | $0.10 | — |
| Output $ / M tokens | $0.30 | — |
| Results tracked | 4 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-4.5 leads
Devstral Small 2505: 38.9 (#166), GPT-4.5: 42.2 (#109)
| Benchmark | Devstral Small 2505 | GPT-4.5 |
|---|---|---|
| SWE-bench Verified (bash only) | 56.4% | — |
| Aider Polyglot | — | 44.9% |
| SciCode | 28.8% | — |
| WeirdML | — | 39.4% |
| LiveBench Coding | — | 75.2% |
| LMArena Coding | — | 1396 |
Agentic & Tool Use Not comparable
Devstral Small 2505: —, GPT-4.5: 27.9 (#97)
| Benchmark | Devstral Small 2505 | GPT-4.5 |
|---|---|---|
| Cybench | — | 17.5% |
Reasoning Devstral Small 2505 leads
Devstral Small 2505: 19.7 (#252), GPT-4.5: 13.9 (#330)
| Benchmark | Devstral Small 2505 | GPT-4.5 |
|---|---|---|
| ARC-AGI-2 | — | 0.8% |
| SimpleBench | — | 34.5% |
| Kagi LLM Benchmark | 37.7% | — |
| ARC-AGI-1 | — | 10.3% |
| CritPt | 0% | — |
| EnigmaEval | — | 3.2% |
| LiveBench Reasoning | — | 71.1% |
| LMArena Hard Prompts | — | 1403 |
| LiveBench Data Analysis | — | 64.3% |
| Epoch Capabilities Index | — | 136.74 |
| ForecastBench | — | 61.7 |
| LiveBench | — | 69% |
Math Not comparable
Devstral Small 2505: —, GPT-4.5: 32.6 (#211)
| Benchmark | Devstral Small 2505 | GPT-4.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 37.8% |
| LiveBench Math | — | 69.3% |
| LMArena Math | — | 1412 |
| MATH Level 5 | — | 78.6% |
Knowledge Not comparable
Devstral Small 2505: —, GPT-4.5: 32.5 (#211)
| Benchmark | Devstral Small 2505 | GPT-4.5 |
|---|---|---|
| GPQA Diamond | — | 68.7% |
| Humanity's Last Exam | — | 5.4% |
| Confabulations | — | 13.6% |
| LMArena Expert | — | 1394 |
Multimodal Not comparable
Devstral Small 2505: —, GPT-4.5: 37.6 (#71)
| Benchmark | Devstral Small 2505 | GPT-4.5 |
|---|---|---|
| LMArena Vision | — | 1195 |
| VPCT | — | 45% |
Multilingual Not comparable
Devstral Small 2505: —, GPT-4.5: 52.5 (#83)
| Benchmark | Devstral Small 2505 | GPT-4.5 |
|---|---|---|
| LMArena Non-English | — | 1413 |
| LMArena Chinese | — | 1421 |
| LMArena French | — | 1418 |
| LMArena German | — | 1457 |
| LMArena Japanese | — | 1416 |
| LMArena Korean | — | 1392 |
| LMArena Russian | — | 1419 |
Instruction Following Not comparable
Devstral Small 2505: —, GPT-4.5: 72.6 (#134)
| Benchmark | Devstral Small 2505 | GPT-4.5 |
|---|---|---|
| LiveBench Instruction Following | — | 72.3% |
| LMArena Instruction Following | — | 1404 |
Long Context Not comparable
Devstral Small 2505: —, GPT-4.5: 40.4 (#155)
| Benchmark | Devstral Small 2505 | GPT-4.5 |
|---|---|---|
| Fiction.LiveBench | — | 63.9% |
| LMArena Longer Query | — | 1406 |
Writing & Preference Not comparable
Devstral Small 2505: —, GPT-4.5: 56.9 (#134)
| Benchmark | Devstral Small 2505 | GPT-4.5 |
|---|---|---|
| LMArena Text | — | 1417 |
| LMArena Creative Writing | — | 1394 |
| Short-Story Creative Writing | — | 75.6% |
| EQ-Bench Creative Writing | — | 1258 |
| LMArena Multi-Turn | — | 1444 |
| LiveBench Language | — | 61.5% |
Frequently asked questions
Is Devstral Small 2505 better than GPT-4.5?
GPT-4.5 is the stronger model overall, scoring 37.2 to 34.3 on the Noometry Index.
Is Devstral Small 2505 or GPT-4.5 better for coding?
GPT-4.5 scores higher on coding benchmarks: 42.2 versus 38.9 in the Noometry coding category.
How many benchmarks do Devstral Small 2505 and GPT-4.5 share?
0 benchmarks have published results for both models. Devstral Small 2505 has 4 scored results on Noometry and GPT-4.5 has 42.