# Devstral 2 vs Gemini 3 Flash Preview

> Gemini 3 Flash Preview has enough public results to be ranked (#40); Devstral 2 does not yet, so treat this comparison as directional.

- Canonical page: https://noometry.com/compare/devstral-2-vs-gemini-3-flash-preview
- Last updated: 2026-10-12
- Shared benchmarks: 2

## Summary

- They share 2 benchmarks with published results for both. Devstral 2 scores higher in 0 categories and Gemini 3 Flash Preview in 1 category; one gap is clear of the uncertainty.
- The widest gap is in coding, where Gemini 3 Flash Preview leads 50.9 to 33.2.
- The biggest single-benchmark swing is SWE-bench Verified (bash only): 53.8% for Devstral 2 and 75.8% for Gemini 3 Flash Preview.
- Devstral 2 is cheaper at $0.40 / $2 per million input/output tokens, against $0.50 / $3 for Gemini 3 Flash Preview.
- Gemini 3 Flash Preview accepts more context: 1.05M tokens versus 262K.
- Devstral 2 has downloadable open weights; the other is API-only.

## Snapshot

| | Devstral 2 | Gemini 3 Flash Preview |
|---|---|---|
| Provider | Mistral AI | Google |
| Noometry Index | 36.1 | 52.3 |
| Rank | — | 40 |
| Context | 262K | 1.05M |
| Input $/M | $0.40 | $0.50 |
| Output $/M | $2 | $3 |
| Weights | Open | Proprietary |

## Coding

- Devstral 2: 33.2 (#261)
- Gemini 3 Flash Preview: 50.9 (#42)

| Benchmark | Devstral 2 | Gemini 3 Flash Preview |
|---|---|---|
| SWE-bench Verified (bash only) | 53.8% | 75.8% |
| LMArena WebDev | 1197 | 1439 |
| SWE-bench Verified | — | 75.4% |
| SWE-bench Multilingual | — | 72.7% |
| GSO | — | 9.8% |
| WeirdML | — | 61.6% |
| LMArena Coding | — | 1460 |
| ALE-Bench | — | 1,367 |

## Agentic & Tool Use

- Devstral 2: —
- Gemini 3 Flash Preview: 38.7 (#29)

| Benchmark | Devstral 2 | Gemini 3 Flash Preview |
|---|---|---|
| Terminal-Bench | — | 64.3% |
| τ²-bench Airline | — | 82.5% |
| τ²-bench Banking | — | 27.3% |
| τ²-bench Retail | — | 76.8% |
| τ²-bench Telecom | — | 91.2% |
| DeepResearch Bench | — | 49.8% |
| BALROG | — | 48.1% |
| GDP.pdf | — | 10% |
| LMArena Search | — | 1198 |
| Vending-Bench 2 | — | 3,635 |

## Reasoning

- Devstral 2: —
- Gemini 3 Flash Preview: 49.2 (#37)

| Benchmark | Devstral 2 | Gemini 3 Flash Preview |
|---|---|---|
| ARC-AGI-2 | — | 33.6% |
| SimpleBench | — | 61.1% |
| NYT Connections (extended) | — | 83.1% |
| ARC-AGI-1 | — | 84.7% |
| Chess Puzzles | — | 40% |
| LMArena Hard Prompts | — | 1465 |
| Mystery Game Puzzles | — | 26% |
| DTBench | — | 89.1% |
| LMCA | — | 43.1% |
| Epoch Capabilities Index | — | 151.8 |
| ForecastBench | — | 58.5 |

## Math

- Devstral 2: —
- Gemini 3 Flash Preview: 51.7 (#55)

| Benchmark | Devstral 2 | Gemini 3 Flash Preview |
|---|---|---|
| FrontierMath (Tiers 1-3) | — | 51.2% |
| FrontierMath Tier 4 | — | 17.1% |
| MathArena Final-Answer Competitions | — | 67.6% |
| OTIS Mock AIME 2024-2025 | — | 95.6% |
| ProofBench | — | 15% |
| LMArena Math | — | 1473 |
| FrontierMath (Feb 2025 set) | — | 35.6% |
| FrontierMath Tier 4 (v1) | — | 4.2% |

## Knowledge

- Devstral 2: —
- Gemini 3 Flash Preview: 58.8 (#33)

| Benchmark | Devstral 2 | Gemini 3 Flash Preview |
|---|---|---|
| GPQA Diamond | — | 89.4% |
| SimpleQA Verified | — | 66.8% |
| Vectara Hallucination Rate | — | 13.5% |
| LMArena Expert | — | 1462 |

## Multimodal

- Devstral 2: —
- Gemini 3 Flash Preview: 45.5 (#16)

| Benchmark | Devstral 2 | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Vision | — | 1285 |
| GeoBench | — | 88% |
| VPCT | — | 72.6% |
| Blueprint-Bench 2 | — | 0% |
| LMArena Document | — | 1413 |

## Multilingual

- Devstral 2: —
- Gemini 3 Flash Preview: 55.7 (#27)

| Benchmark | Devstral 2 | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Non-English | — | 1458 |
| LMArena Chinese | — | 1511 |
| LMArena French | — | 1477 |
| LMArena German | — | 1497 |
| LMArena Japanese | — | 1489 |
| LMArena Korean | — | 1443 |
| LMArena Russian | — | 1480 |
| LMArena Spanish | — | 1469 |

## Instruction Following

- Devstral 2: —
- Gemini 3 Flash Preview: 75.7 (#56)

| Benchmark | Devstral 2 | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Instruction Following | — | 1437 |

## Long Context

- Devstral 2: —
- Gemini 3 Flash Preview: 44.4 (#67)

| Benchmark | Devstral 2 | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Longer Query | — | 1452 |

## Writing & Preference

- Devstral 2: —
- Gemini 3 Flash Preview: 65.5 (#45)

| Benchmark | Devstral 2 | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Text | — | 1466 |
| LMArena Creative Writing | — | 1457 |
| LMArena Multi-Turn | — | 1471 |

## FAQ

### Is Devstral 2 better than Gemini 3 Flash Preview?

Gemini 3 Flash Preview has enough public results to be ranked (#40); Devstral 2 does not yet, so treat this comparison as directional.

### Which is cheaper, Devstral 2 or Gemini 3 Flash Preview?

Devstral 2 is cheaper. It lists at $0.40 per million input tokens and $2 per million output tokens; Gemini 3 Flash Preview lists at $0.50 and $3.

### Is Devstral 2 or Gemini 3 Flash Preview better for coding?

Gemini 3 Flash Preview scores higher on coding benchmarks: 50.9 versus 33.2 in the Noometry coding category.

### Which has the bigger context window?

Gemini 3 Flash Preview does, with 1.05M tokens against 262K.

### How many benchmarks do Devstral 2 and Gemini 3 Flash Preview share?

2 benchmarks have published results for both models. Devstral 2 has 2 scored results on Noometry and Gemini 3 Flash Preview has 59.