Model comparison
Codestral vs Step 3.5 Flash
Step 3.5 Flash is the stronger model overall, scoring 42.3 to 30.6 on the Noometry Index.
Last verified . 0 shared benchmarks.
Mistral AI
StepFunSummary
- The widest gap is in coding, where Step 3.5 Flash leads 42.4 to 27.3.
- Step 3.5 Flash is cheaper at $0.10 / $0.30 per million input/output tokens, against $0.30 / $0.90 for Codestral.
- Step 3.5 Flash has downloadable open weights; the other is API-only.
Side by side
| Codestral | Step 3.5 Flash | |
|---|---|---|
| Provider | Mistral AI | StepFun |
| Noometry Index | 30.6 | 42.3 |
| Released | 2024-05-29 | 2026-01-29 |
| Weights | Proprietary | Open |
| Context window | 256K | 256K |
| Max output | 8K | 256K |
| Input $ / M tokens | $0.30 | $0.10 |
| Output $ / M tokens | $0.90 | $0.30 |
| Results tracked | 7 | 19 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 3.5 Flash leads
Codestral: 27.3 (#321), Step 3.5 Flash: 42.4 (#105)
| Benchmark | Codestral | Step 3.5 Flash |
|---|---|---|
| Aider Polyglot | 11.1% | — |
| BigCodeBench Instruct | 41.8% | — |
| LMArena Coding | — | 1436 |
| BigCodeBench Complete | 52.5% | — |
| ALE-Bench | 137.78 | — |
| HumanEval+ | 73.8% | — |
| MBPP+ | 61.9% | — |
Reasoning Step 3.5 Flash leads
Codestral: 19.8 (#251), Step 3.5 Flash: 22.2 (#202)
| Benchmark | Codestral | Step 3.5 Flash |
|---|---|---|
| Kagi LLM Benchmark | 32.5% | — |
| NYT Connections (extended) | — | 28.4% |
| LMArena Hard Prompts | — | 1411 |
Math Not comparable
Codestral: —, Step 3.5 Flash: 42.6 (#84)
| Benchmark | Codestral | Step 3.5 Flash |
|---|---|---|
| MathArena Final-Answer Competitions | — | 66.8% |
| LMArena Math | — | 1408 |
Knowledge Not comparable
Codestral: —, Step 3.5 Flash: 39.6 (#132)
| Benchmark | Codestral | Step 3.5 Flash |
|---|---|---|
| LMArena Expert | — | 1421 |
Multilingual Not comparable
Codestral: —, Step 3.5 Flash: 50.5 (#119)
| Benchmark | Codestral | Step 3.5 Flash |
|---|---|---|
| LMArena Non-English | — | 1385 |
| LMArena Chinese | — | 1447 |
| LMArena French | — | 1421 |
| LMArena German | — | 1405 |
| LMArena Japanese | — | 1354 |
| LMArena Korean | — | 1352 |
| LMArena Russian | — | 1385 |
| LMArena Spanish | — | 1419 |
Instruction Following Not comparable
Codestral: —, Step 3.5 Flash: 73.1 (#124)
| Benchmark | Codestral | Step 3.5 Flash |
|---|---|---|
| LMArena Instruction Following | — | 1385 |
Long Context Not comparable
Codestral: —, Step 3.5 Flash: 42.8 (#117)
| Benchmark | Codestral | Step 3.5 Flash |
|---|---|---|
| LMArena Longer Query | — | 1402 |
Writing & Preference Not comparable
Codestral: —, Step 3.5 Flash: 58.8 (#113)
| Benchmark | Codestral | Step 3.5 Flash |
|---|---|---|
| LMArena Text | — | 1403 |
| LMArena Creative Writing | — | 1357 |
| LMArena Multi-Turn | — | 1405 |
Frequently asked questions
Is Codestral better than Step 3.5 Flash?
Step 3.5 Flash is the stronger model overall, scoring 42.3 to 30.6 on the Noometry Index.
Which is cheaper, Codestral or Step 3.5 Flash?
Step 3.5 Flash is cheaper. It lists at $0.10 per million input tokens and $0.30 per million output tokens; Codestral lists at $0.30 and $0.90.
Is Codestral or Step 3.5 Flash better for coding?
Step 3.5 Flash scores higher on coding benchmarks: 42.4 versus 27.3 in the Noometry coding category.
Which has the bigger context window?
Both accept 256K tokens.
How many benchmarks do Codestral and Step 3.5 Flash share?
0 benchmarks have published results for both models. Codestral has 7 scored results on Noometry and Step 3.5 Flash has 19.