# DeepSeek V4 Flash 0731 vs Qwen3.8 Max

> Side-by-side benchmark comparison for DeepSeek V4 Flash 0731 and Qwen3.8 Max across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/deepseek-v4-flash-0731-vs-qwen3-8-max
- Last updated: September 27, 2026

- Shared sourced benchmarks: 19
- HTML indexing: indexable
- Ranking lane: BenchAlign v5.7

## Quick Verdict

Pick Qwen3.8 Max if you want the stronger benchmark profile. DeepSeek V4 Flash 0731 only makes more sense when its price, context window, or workload-specific category wins matter more than the overall score.

## Summary

- Qwen3.8 Max leads overall 71.8 to 57.
- The clearest category separation is in reasoning, where the averages are 57 for DeepSeek V4 Flash 0731 and 87.6 for Qwen3.8 Max.
- The biggest single benchmark swing is Agents' Last Exam in Agentic, with scores of 25.2% and 52.4%.
- Qwen3.8 Max is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.

## Model Snapshot

| Property | DeepSeek V4 Flash 0731 | Qwen3.8 Max |
|----------|----------|----------|
| Creator | DeepSeek | Alibaba |
| Type | Open Weight | Open Weight |
| Reasoning | Reasoning | Reasoning |
| Context | 1M | 1M |
| Overall Score | null | 71.8 |
| Benchmarks Covered | 42 | 60 |
| Pricing (input/output) | $0.14 / $0.28 | Pricing unavailable / Pricing unavailable |

## Category Breakdown

### Agentic

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: Coming soon
- Qwen3.8 Max public-lane score: 65.6 (Supported · #10/105)

| Benchmark | DeepSeek V4 Flash 0731 | Qwen3.8 Max | Winner |
|-----------|-----------|-----------|--------|
| Terminal-Bench 2.0 | 56.9% | Coming soon | Coming soon |
| BrowseComp | 73.2% | Coming soon | Coming soon |
| HLE w/ tools | 45.1% | 56.2% | Qwen3.8 Max |
| MCP Atlas | 69% | Coming soon | Coming soon |
| Toolathlon | 47.8% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | 82.7% | 86.6% | Qwen3.8 Max |
| CyberGym | 76.7% | Coming soon | Coming soon |
| Toolathlon-Verified | 70.3% | 72.5% | Qwen3.8 Max |
| Agents' Last Exam | 25.2% | 52.4% | Qwen3.8 Max |
| AutomationBench | 25.1% | 27.3% | Qwen3.8 Max |
| Terminal-Bench 2.1 (Vals) | 67.0% | 67.4% | Qwen3.8 Max |
| CoWorkBench | Coming soon | 74.8% | Coming soon |
| JobBench | Coming soon | 53.4% | Coming soon |
| skillsBench | Coming soon | 70.2% | Coming soon |
| WideResearch | Coming soon | 81.9% | Coming soon |
| OSWorld-Verified | Coming soon | 86.1% | Coming soon |
| OSWorld 2.0 | Coming soon | 19.4% | Coming soon |
| WebArena-Verified | Coming soon | 66.8% | Coming soon |
| AndroidWorld | Coming soon | 85.3% | Coming soon |
| MobileWorld | Coming soon | 77.8% | Coming soon |

### Coding

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: Coming soon
- Qwen3.8 Max public-lane score: 56 (Supported · #24/135)

| Benchmark | DeepSeek V4 Flash 0731 | Qwen3.8 Max | Winner |
|-----------|-----------|-----------|--------|
| LiveCodeBench Pass@1-COT | 91.6% | Coming soon | Coming soon |
| Codeforces | 3052.0 | Coming soon | Coming soon |
| SWE-bench Verified | 79% | Coming soon | Coming soon |
| SWE-bench Pro | 52.6% | 67.7% | Qwen3.8 Max |
| SWE Multilingual | 73.3% | Coming soon | Coming soon |
| Terminal-Bench 2.0 | 56.9% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | 82.7% | 86.6% | Qwen3.8 Max |
| NL2Repo | 54.2% | 55.9% | Qwen3.8 Max |
| DeepSWE | 54.4% | 56.6% | Qwen3.8 Max |
| DSBench-FullStack | 68.7% | Coming soon | Coming soon |
| DSBench-Hard | 59.6% | Coming soon | Coming soon |
| VulcanBench v3 | 88.4% | 81.2% | DeepSeek V4 Flash 0731 |
| OpenHarmony Bench | 53.8% | 60.8% | Qwen3.8 Max |
| LiveCodeBench (Vals) | 87.3% | 87.9% | Qwen3.8 Max |
| SWE-bench (Vals) | 88.8% | 85.6% | DeepSeek V4 Flash 0731 |
| FrontierSWE | Coming soon | 73.5% | Coming soon |
| MLS-Bench Lite | Coming soon | 41.0% | Coming soon |
| PaperBench | Coming soon | 93.0% | Coming soon |
| FrontierSWE v2 | Coming soon | 15.8% | Coming soon |

### Reasoning

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: 57 (Unranked · 4 rankable rows)
- Qwen3.8 Max public-lane score: 87.6 (Unranked · 2 rankable rows)

| Benchmark | DeepSeek V4 Flash 0731 | Qwen3.8 Max | Winner |
|-----------|-----------|-----------|--------|
| MRCR 1M | 78.7% | Coming soon | Coming soon |
| CorpusQA 1M | 60.5% | Coming soon | Coming soon |
| ARC-AGI-1 | 89.00% | Coming soon | Coming soon |
| ARC-AGI-2 | 61.4% | Coming soon | Coming soon |
| MRCRv2 | Coming soon | 92.9% | Coming soon |
| LongBench v2 | Coming soon | 66.3% | Coming soon |

### Multimodal & Grounded

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: Coming soon
- Qwen3.8 Max public-lane score: 87.4 (#6/50)

| Benchmark | DeepSeek V4 Flash 0731 | Qwen3.8 Max | Winner |
|-----------|-----------|-----------|--------|
| MMMU-Pro | Coming soon | 82.3% | Coming soon |
| MathVision | Coming soon | 95.2% | Coming soon |
| MathVision w/ Python | Coming soon | 97.7% | Coming soon |
| BabyVision | Coming soon | 82.0% | Coming soon |
| BabyVision w/ Python | Coming soon | 91.3% | Coming soon |
| ZeroBench | Coming soon | 24.0% | Coming soon |
| ZeroBench w/ Python | Coming soon | 49.0% | Coming soon |
| MedXpertQA (MM) | Coming soon | 80.4% | Coming soon |
| ScreenSpot Pro | Coming soon | 84.5% | Coming soon |
| Vision2Web | Coming soon | 69.0% | Coming soon |
| CharXiv w/o tools | Coming soon | 88.4% | Coming soon |
| CharXiv | Coming soon | 93.5% | Coming soon |
| OmniDocBench 1.5 | Coming soon | 92.1% | Coming soon |
| OCRBench V2 | Coming soon | 74.2% | Coming soon |
| CC-OCR | Coming soon | 79.6% | Coming soon |
| RealWorldQA | Coming soon | 88.0% | Coming soon |
| ERQA | Coming soon | 77.8% | Coming soon |
| SimpleVQA | Coming soon | 75.0% | Coming soon |
| PerceptionBench | Coming soon | 63.5% | Coming soon |
| Video-MME (with subtitle) | Coming soon | 90.4% | Coming soon |
| VideoMMMU | Coming soon | 88.7% | Coming soon |
| MMVU | Coming soon | 82.4% | Coming soon |
| MLVU (M-Avg) | Coming soon | 90.8% | Coming soon |
| LVBench | Coming soon | 81.8% | Coming soon |

### Knowledge

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: Coming soon
- Qwen3.8 Max public-lane score: 66.7 (Supported · #19/158)

| Benchmark | DeepSeek V4 Flash 0731 | Qwen3.8 Max | Winner |
|-----------|-----------|-----------|--------|
| MMLU-Pro | 86.2% | Coming soon | Coming soon |
| SimpleQA | 34.1% | Coming soon | Coming soon |
| Chinese-SimpleQA | 78.9% | Coming soon | Coming soon |
| GPQA | 88.1% | 92.6% | Qwen3.8 Max |
| GPQA-D | 88.1% | 92.6% | Qwen3.8 Max |
| HLE | 34.8% | 43.6% | Qwen3.8 Max |
| GPQA Diamond (Vals) | 89.9% | 93.7% | Qwen3.8 Max |
| MMLU-Pro (Vals) | 86.2% | 88.6% | Qwen3.8 Max |
| HLE w/o tools | Coming soon | 43.6% | Coming soon |

### Instruction Following

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: Coming soon
- Qwen3.8 Max public-lane score: 90.5 (#16/124)

| Benchmark | DeepSeek V4 Flash 0731 | Qwen3.8 Max | Winner |
|-----------|-----------|-----------|--------|
| IFBench | Coming soon | 82.8% | Coming soon |

### Mathematics

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: 79.9 (Unranked · 4 rankable rows)
- Qwen3.8 Max public-lane score: Coming soon

| Benchmark | DeepSeek V4 Flash 0731 | Qwen3.8 Max | Winner |
|-----------|-----------|-----------|--------|
| HMMT Feb 2026 | 94.8% | Coming soon | Coming soon |
| IMOAnswerBench | 88.4% | Coming soon | Coming soon |
| Apex | 33.0% | Coming soon | Coming soon |
| Apex Shortlist | 85.7% | Coming soon | Coming soon |

## FAQ

### Which is better overall, DeepSeek V4 Flash 0731 or Qwen3.8 Max?

Qwen3.8 Max is ahead overall on BenchLM right now.

### Where is the biggest gap between DeepSeek V4 Flash 0731 and Qwen3.8 Max?

The widest category gap is in reasoning, where the averages are 57 for DeepSeek V4 Flash 0731 and 87.6 for Qwen3.8 Max.

### How many benchmarks does DeepSeek V4 Flash 0731 cover on BenchLM?

DeepSeek V4 Flash 0731 currently has 42 sourced benchmark scores on BenchLM.

### How many benchmarks does Qwen3.8 Max cover on BenchLM?

Qwen3.8 Max currently has 60 sourced benchmark scores on BenchLM.

## Related Comparisons

- [DeepSeek V4 Flash 0731 vs Qwen3.8 Max Preview](/compare/deepseek-v4-flash-0731-vs-qwen3-8-max-preview)
- [Qwen3.8 Max vs Qwen3.8 Max Preview](/compare/qwen3-8-max-vs-qwen3-8-max-preview)
- [DeepSeek V4 Flash 0731 vs GPT-6 Astra](/compare/deepseek-v4-flash-0731-vs-gpt-6-astra)
- [Qwen3.8 Max vs GPT-6 Astra](/compare/gpt-6-astra-vs-qwen3-8-max)

## Explore More

- [DeepSeek V4 Flash 0731 profile](/models/deepseek-v4-flash-0731)
- [Qwen3.8 Max profile](/models/qwen3-8-max)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)