Model comparison
Devstral Small 2505 vs Trinity Large Thinking
Trinity Large Thinking is the stronger model overall, scoring 38.6 to 34.3 on the Noometry Index. Devstral Small 2505 costs 2.6× less per token, which makes it the better buy when Trinity Large Thinking's lead doesn't matter for your workload.
Last verified . 2 shared benchmarks.
Mistral AI
Arcee AISummary
- They share 2 benchmarks with published results for both. Devstral Small 2505 scores higher in 2 categories and Trinity Large Thinking in 0 categories; 2 gaps are clear of the uncertainty.
- The widest gap is in coding, where Devstral Small 2505 leads 38.9 to 34.1.
- The biggest single-benchmark swing is SciCode: 28.8% for Devstral Small 2505 and 36.1% for Trinity Large Thinking.
- Devstral Small 2505 is cheaper at $0.10 / $0.30 per million input/output tokens, against $0.25 / $0.80 for Trinity Large Thinking.
- Trinity Large Thinking accepts more context: 262K tokens versus 128K.
Side by side
| Devstral Small 2505 | Trinity Large Thinking | |
|---|---|---|
| Provider | Mistral AI | Arcee AI |
| Noometry Index | 34.3 | 38.6 |
| Released | 2025-05-07 | 2026-04-01 |
| Weights | Open | Open |
| Context window | 128K | 262K |
| Max output | 128K | 80K |
| Input $ / M tokens | $0.10 | $0.25 |
| Output $ / M tokens | $0.30 | $0.80 |
| Results tracked | 4 | 24 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Devstral Small 2505 leads
Devstral Small 2505: 38.9 (#166), Trinity Large Thinking: 34.1 (#244)
| Benchmark | Devstral Small 2505 | Trinity Large Thinking |
|---|---|---|
| SciCode | 28.8% | 36.1% |
| SWE-bench Verified (bash only) | 56.4% | — |
| LMArena WebDev | — | 1238 |
| LMArena Coding | — | 1381 |
Reasoning Devstral Small 2505 leads
Devstral Small 2505: 19.7 (#252), Trinity Large Thinking: 16.9 (#298)
| Benchmark | Devstral Small 2505 | Trinity Large Thinking |
|---|---|---|
| CritPt | 0% | 0.9% |
| Kagi LLM Benchmark | 37.7% | — |
| NYT Connections (extended) | — | 16.5% |
| Thematic Generalization | — | 41.6% |
| LMArena Hard Prompts | — | 1350 |
| Surface Evolver Bench | — | 15.6% |
Math Not comparable
Devstral Small 2505: —, Trinity Large Thinking: 37.6 (#149)
| Benchmark | Devstral Small 2505 | Trinity Large Thinking |
|---|---|---|
| LMArena Math | — | 1366 |
Knowledge Not comparable
Devstral Small 2505: —, Trinity Large Thinking: 40.9 (#113)
| Benchmark | Devstral Small 2505 | Trinity Large Thinking |
|---|---|---|
| Vectara Hallucination Rate | — | 6.9% |
| LMArena Expert | — | 1360 |
Multilingual Not comparable
Devstral Small 2505: —, Trinity Large Thinking: 46.2 (#160)
| Benchmark | Devstral Small 2505 | Trinity Large Thinking |
|---|---|---|
| LMArena Non-English | — | 1325 |
| LMArena Chinese | — | 1373 |
| LMArena French | — | 1374 |
| LMArena German | — | 1356 |
| LMArena Japanese | — | 1311 |
| LMArena Korean | — | 1306 |
| LMArena Russian | — | 1337 |
| LMArena Spanish | — | 1357 |
Instruction Following Not comparable
Devstral Small 2505: —, Trinity Large Thinking: 70.5 (#162)
| Benchmark | Devstral Small 2505 | Trinity Large Thinking |
|---|---|---|
| LMArena Instruction Following | — | 1334 |
Long Context Not comparable
Devstral Small 2505: —, Trinity Large Thinking: 41.3 (#144)
| Benchmark | Devstral Small 2505 | Trinity Large Thinking |
|---|---|---|
| LMArena Longer Query | — | 1355 |
Writing & Preference Not comparable
Devstral Small 2505: —, Trinity Large Thinking: 53.8 (#158)
| Benchmark | Devstral Small 2505 | Trinity Large Thinking |
|---|---|---|
| LMArena Text | — | 1340 |
| LMArena Creative Writing | — | 1320 |
| LMArena Multi-Turn | — | 1342 |
Frequently asked questions
Is Devstral Small 2505 better than Trinity Large Thinking?
Trinity Large Thinking is the stronger model overall, scoring 38.6 to 34.3 on the Noometry Index. Devstral Small 2505 costs 2.6× less per token, which makes it the better buy when Trinity Large Thinking's lead doesn't matter for your workload.
Which is cheaper, Devstral Small 2505 or Trinity Large Thinking?
Devstral Small 2505 is cheaper. It lists at $0.10 per million input tokens and $0.30 per million output tokens; Trinity Large Thinking lists at $0.25 and $0.80.
Is Devstral Small 2505 or Trinity Large Thinking better for coding?
Devstral Small 2505 scores higher on coding benchmarks: 38.9 versus 34.1 in the Noometry coding category.
Which has the bigger context window?
Trinity Large Thinking does, with 262K tokens against 128K.
How many benchmarks do Devstral Small 2505 and Trinity Large Thinking share?
2 benchmarks have published results for both models. Devstral Small 2505 has 4 scored results on Noometry and Trinity Large Thinking has 24.