# Deepseek Coder v2 vs Mistral Small

> Deepseek Coder v2 is the stronger model overall, scoring 35.3 to 33.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-coder-v2-vs-mistral-small
- Last updated: 2026-10-10
- Shared benchmarks: 19

## Summary

- They share 19 benchmarks with published results for both. Deepseek Coder v2 scores higher in 4 categories and Mistral Small in 4 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Deepseek Coder v2 leads 32.3 to 15.4.
- The biggest single-benchmark swing is BigCodeBench Complete: 59.7% for Deepseek Coder v2 and 46.6% for Mistral Small.

## Snapshot

| | Deepseek Coder v2 | Mistral Small |
|---|---|---|
| Provider | DeepSeek | Mistral AI |
| Noometry Index | 35.3 | 33.1 |
| Rank | 217 | 238 |
| Context | — | 256K |
| Input $/M | — | $0.15 |
| Output $/M | — | $0.60 |
| Weights | Open | Open |

## Coding

- Deepseek Coder v2: 37.5 (#178)
- Mistral Small: 33.5 (#242)

| Benchmark | Deepseek Coder v2 | Mistral Small |
|---|---|---|
| BigCodeBench Instruct | 48.2% | 36.1% |
| LMArena Coding | 1251 | 1363 |
| BigCodeBench Complete | 59.7% | 46.6% |
| SciCode | — | 26.5% |
| LiveBench Coding | — | 36.2% |
| ALE-Bench | — | 497.62 |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |

## Agentic & Tool Use

- Deepseek Coder v2: —
- Mistral Small: 27.9 (#87)

| Benchmark | Deepseek Coder v2 | Mistral Small |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 37.1% |

## Reasoning

- Deepseek Coder v2: 22.6 (#175)
- Mistral Small: 19.3 (#238)

| Benchmark | Deepseek Coder v2 | Mistral Small |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1336 |
| Kagi LLM Benchmark | — | 37.8% |
| CritPt | — | 0% |
| LiveBench Reasoning | — | 44.8% |
| DTBench | — | 70.9% |
| LiveBench Data Analysis | — | 53.7% |
| LMCA | — | 20.6% |
| LiveBench | — | 44% |
| WinoGrande | 83.7% | — |

## Math

- Deepseek Coder v2: 32.3 (#188)
- Mistral Small: 15.4 (#287)

| Benchmark | Deepseek Coder v2 | Mistral Small |
|---|---|---|
| LMArena Math | 1241 | 1341 |
| OTIS Mock AIME 2024-2025 | — | 5.8% |
| LiveBench Math | — | 39.9% |
| MATH Level 5 | — | 46.8% |
| GSM8K | 94.5% | — |

## Knowledge

- Deepseek Coder v2: 31.7 (#209)
- Mistral Small: 30.9 (#214)

| Benchmark | Deepseek Coder v2 | Mistral Small |
|---|---|---|
| LMArena Expert | 1181 | 1293 |
| GPQA Diamond | — | 47.5% |
| Vectara Hallucination Rate | — | 5.1% |
| ARC (AI2) Challenge | 64.3% | — |
| MMLU | — | 68.7% |

## Multimodal

- Deepseek Coder v2: —
- Mistral Small: 32.9 (#91)

| Benchmark | Deepseek Coder v2 | Mistral Small |
|---|---|---|
| LMArena Vision | — | 1142 |

## Multilingual

- Deepseek Coder v2: 36.4 (#235)
- Mistral Small: 45.7 (#164)

| Benchmark | Deepseek Coder v2 | Mistral Small |
|---|---|---|
| LMArena Non-English | 1182 | 1315 |
| LMArena Chinese | 1201 | 1342 |
| LMArena French | 1184 | 1337 |
| LMArena German | 1164 | 1341 |
| LMArena Japanese | 1126 | 1275 |
| LMArena Korean | 1104 | 1260 |
| LMArena Russian | 1188 | 1324 |
| LMArena Spanish | 1153 | 1346 |

## Instruction Following

- Deepseek Coder v2: 61.9 (#237)
- Mistral Small: 66.6 (#204)

| Benchmark | Deepseek Coder v2 | Mistral Small |
|---|---|---|
| LMArena Instruction Following | 1180 | 1310 |
| LiveBench Instruction Following | — | 63.7% |

## Long Context

- Deepseek Coder v2: 37.0 (#219)
- Mistral Small: 40.4 (#151)

| Benchmark | Deepseek Coder v2 | Mistral Small |
|---|---|---|
| LMArena Longer Query | 1219 | 1327 |

## Writing & Preference

- Deepseek Coder v2: 38.4 (#249)
- Mistral Small: 52.6 (#165)

| Benchmark | Deepseek Coder v2 | Mistral Small |
|---|---|---|
| LMArena Text | 1191 | 1338 |
| LMArena Creative Writing | 1120 | 1304 |
| LMArena Multi-Turn | 1176 | 1344 |
| LiveBench Language | — | 30.5% |

## FAQ

### Is Deepseek Coder v2 better than Mistral Small?

Deepseek Coder v2 is the stronger model overall, scoring 35.3 to 33.1 on the Noometry Index.

### Is Deepseek Coder v2 or Mistral Small better for coding?

Deepseek Coder v2 scores higher on coding benchmarks: 37.5 versus 33.5 in the Noometry coding category.

### How many benchmarks do Deepseek Coder v2 and Mistral Small share?

19 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Mistral Small has 39.