# DeepSeek V4 Flash vs Qwen2.5-Max

> DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 40.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-flash-vs-qwen2-5-max
- Last updated: 2026-10-10
- Shared benchmarks: 18

## Summary

- They share 18 benchmarks with published results for both. DeepSeek V4 Flash scores higher in 8 categories and Qwen2.5-Max in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Flash leads 53.7 to 24.6.
- DeepSeek V4 Flash has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek V4 Flash | Qwen2.5-Max |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 53.6 | 40.1 |
| Rank | 32 | 142 |
| Context | 1M | — |
| Input $/M | $0.15 | — |
| Output $/M | $0.60 | — |
| Weights | Open | Proprietary |

## Coding

- DeepSeek V4 Flash: 47.6 (#56)
- Qwen2.5-Max: 41.2 (#118)

| Benchmark | DeepSeek V4 Flash | Qwen2.5-Max |
|---|---|---|
| LMArena Coding | 1457 | 1359 |
| FrontierCode | 18.8% | — |
| LMArena WebDev | 1581 | — |
| SciCode | 49.9% | — |
| WeirdML | 63% | — |
| LiveBench Coding | — | 64.4% |
| ALE-Bench | 1,306 | — |

## Reasoning

- DeepSeek V4 Flash: 53.7 (#25)
- Qwen2.5-Max: 24.6 (#144)

| Benchmark | DeepSeek V4 Flash | Qwen2.5-Max |
|---|---|---|
| LMArena Hard Prompts | 1443 | 1360 |
| Epoch Capabilities Index | 154.51 | 132.54 |
| ARC-AGI-2 | 61.4% | — |
| SimpleBench | 61.1% | — |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 89.6% | — |
| ARC-AGI-1 | 89% | — |
| CritPt | 16.6% | — |
| Chess Puzzles | 33% | — |
| LiveBench Reasoning | — | 51.4% |
| Mystery Game Puzzles | 34% | — |
| DTBench | 90.9% | — |
| LiveBench Data Analysis | — | 67.9% |
| LMCA | 41.7% | — |
| LiveBench | — | 62.3% |

## Math

- DeepSeek V4 Flash: 60.0 (#31)
- Qwen2.5-Max: 34.3 (#162)

| Benchmark | DeepSeek V4 Flash | Qwen2.5-Max |
|---|---|---|
| LMArena Math | 1427 | 1369 |
| FrontierMath (Tiers 1-3) | 57.5% | — |
| FrontierMath Tier 4 | 24.4% | — |
| MathArena Final-Answer Competitions | 76.5% | — |
| OTIS Mock AIME 2024-2025 | 94.4% | — |
| ProofBench | 56% | — |
| LiveBench Math | — | 58.4% |

## Knowledge

- DeepSeek V4 Flash: 55.5 (#43)
- Qwen2.5-Max: 34.8 (#181)

| Benchmark | DeepSeek V4 Flash | Qwen2.5-Max |
|---|---|---|
| LMArena Expert | 1441 | 1337 |
| GPQA Diamond | 91% | — |
| SimpleQA Verified | 33.6% | — |
| Confabulations | — | 21.8% |

## Multilingual

- DeepSeek V4 Flash: 53.2 (#71)
- Qwen2.5-Max: 48.3 (#140)

| Benchmark | DeepSeek V4 Flash | Qwen2.5-Max |
|---|---|---|
| LMArena Non-English | 1420 | 1352 |
| LMArena Chinese | 1467 | 1382 |
| LMArena French | 1438 | 1396 |
| LMArena German | 1418 | 1350 |
| LMArena Japanese | 1403 | 1299 |
| LMArena Korean | 1383 | 1304 |
| LMArena Russian | 1428 | 1353 |
| LMArena Spanish | 1437 | 1377 |

## Instruction Following

- DeepSeek V4 Flash: 75.1 (#76)
- Qwen2.5-Max: 71.5 (#147)

| Benchmark | DeepSeek V4 Flash | Qwen2.5-Max |
|---|---|---|
| LMArena Instruction Following | 1421 | 1335 |
| LiveBench Instruction Following | — | 75.3% |

## Long Context

- DeepSeek V4 Flash: 43.8 (#81)
- Qwen2.5-Max: 41.3 (#137)

| Benchmark | DeepSeek V4 Flash | Qwen2.5-Max |
|---|---|---|
| LMArena Longer Query | 1435 | 1358 |

## Writing & Preference

- DeepSeek V4 Flash: 64.0 (#59)
- Qwen2.5-Max: 55.6 (#141)

| Benchmark | DeepSeek V4 Flash | Qwen2.5-Max |
|---|---|---|
| LMArena Text | 1432 | 1367 |
| LMArena Creative Writing | 1403 | 1339 |
| LMArena Multi-Turn | 1449 | 1364 |
| Short-Story Creative Writing | — | 72.9% |
| EQ-Bench Creative Writing | 1559 | — |
| LiveBench Language | — | 56.3% |

## FAQ

### Is DeepSeek V4 Flash better than Qwen2.5-Max?

DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 40.1 on the Noometry Index.

### Is DeepSeek V4 Flash or Qwen2.5-Max better for coding?

DeepSeek V4 Flash scores higher on coding benchmarks: 47.6 versus 41.2 in the Noometry coding category.

### How many benchmarks do DeepSeek V4 Flash and Qwen2.5-Max share?

18 benchmarks have published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and Qwen2.5-Max has 27.