# DeepSeek V4 Pro vs Qwen3.8 Max

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

- Canonical page: https://benchlm.ai/compare/deepseek-v4-pro-vs-qwen3-8-max
- Last updated: November 21, 2026

- Shared sourced benchmarks: 21
- HTML indexing: indexable
- Ranking lane: BenchAlign v6.0

## Quick Verdict

Pick Qwen3.8 Max if you want the stronger benchmark profile. DeepSeek V4 Pro 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 65.3 to 61.6.
- The clearest category separation is in coding, where the averages are 59.6 for DeepSeek V4 Pro and 67.5 for Qwen3.8 Max.
- The biggest single benchmark swing is Agents' Last Exam in Agentic, with scores of 25.7% 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 Pro | Qwen3.8 Max |
|----------|----------|----------|
| Creator | DeepSeek | Alibaba |
| Type | Open Weight | Open Weight |
| Reasoning | Reasoning | Reasoning |
| Context | 1M | 1M |
| Overall Score | 61.6 | 65.3 |
| Benchmarks Covered | 46 | 70 |
| Pricing (input/output) | $1.32 / $3.96 | Pricing unavailable / Pricing unavailable |

## Category Breakdown

### Agentic

- Winner: Qwen3.8 Max
- DeepSeek V4 Pro public-lane score: 60 (Supported · #55/182)
- Qwen3.8 Max public-lane score: 64.8 (Supported · #38/182)

| Benchmark | DeepSeek V4 Pro | Qwen3.8 Max | Winner |
|-----------|-----------|-----------|--------|
| Gert Labs | 52.74% | 55.66% | Qwen3.8 Max |
| Terminal-Bench 4.1 | 6.70% | Coming soon | Coming soon |
| Finch | 67.4% | Coming soon | Coming soon |
| Terminal-Bench 2.0 | 67.9% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | 87.9% | 86.6% | DeepSeek V4 Pro |
| BrowseComp | 83.4% | Coming soon | Coming soon |
| HLE w/ tools | 60.0% | 56.2% | DeepSeek V4 Pro |
| MCP Atlas | 73.6% | Coming soon | Coming soon |
| Toolathlon | 51.8% | Coming soon | Coming soon |
| CyberGym | 83.3% | Coming soon | Coming soon |
| Toolathlon-Verified | 74.1% | 72.5% | DeepSeek V4 Pro |
| Agents' Last Exam | 25.7% | 52.4% | Qwen3.8 Max |
| AutomationBench | 31.8% | 27.3% | DeepSeek V4 Pro |
| Terminal-Bench 2.1 (Vals) | 54.7% | 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 |
| Hermes Index | Coming soon | 36.2% | Coming soon |
| Hermes Bench | Coming soon | 70.5% | Coming soon |
| MazeBench | Coming soon | 0% | Coming soon |
| CWE-bench | Coming soon | 45.4% | Coming soon |

### Coding

- Winner: Qwen3.8 Max
- DeepSeek V4 Pro public-lane score: 59.6 (Supported · #56/180)
- Qwen3.8 Max public-lane score: 67.5 (Supported · #30/180)

| Benchmark | DeepSeek V4 Pro | Qwen3.8 Max | Winner |
|-----------|-----------|-----------|--------|
| LiveCodeBench Pass@1-COT | 93.5% | Coming soon | Coming soon |
| Codeforces | 3206.0 | Coming soon | Coming soon |
| SWE-bench Verified | 80.6% | Coming soon | Coming soon |
| SWE-bench Pro | 55.4% | 67.7% | Qwen3.8 Max |
| SWE Multilingual | 76.2% | Coming soon | Coming soon |
| Terminal-Bench 2.0 | 67.9% | Coming soon | Coming soon |
| Vibe Code Bench | 49.93% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | 87.9% | 86.6% | DeepSeek V4 Pro |
| NL2Repo | 61.5% | 55.9% | DeepSeek V4 Pro |
| DeepSWE | 62.7% | 56.6% | DeepSeek V4 Pro |
| DSBench-FullStack | 71.1% | Coming soon | Coming soon |
| DSBench-Hard | 67.2% | Coming soon | Coming soon |
| OpenHarmony Bench v1.0 | 59.0% | 60.8% | Qwen3.8 Max |
| LiveCodeBench (Vals) | 87.5% | 87.9% | Qwen3.8 Max |
| SWE-bench (Vals) | 96.4% | 85.6% | DeepSeek V4 Pro |
| FrontierSWE | Coming soon | 73.5% | Coming soon |
| MLS-Bench Lite | Coming soon | 41.0% | Coming soon |
| PaperBench | Coming soon | 93.0% | Coming soon |
| VulcanBench v3 | Coming soon | 81.2% | Coming soon |
| FrontierSWE v2 | Coming soon | 15.8% | Coming soon |
| Bug Hunt Bench | Coming soon | 25.7 fixes | Coming soon |
| ScarfBench | Coming soon | 63.7% | Coming soon |
| OpenHarmony Bench v1.2 | Coming soon | 67.77% | Coming soon |

### Reasoning

- Winner: Tie
- DeepSeek V4 Pro public-lane score: 64.6 (Supported · #32/161)
- Qwen3.8 Max public-lane score: 64.7 (Supported · #31/161)

| Benchmark | DeepSeek V4 Pro | Qwen3.8 Max | Winner |
|-----------|-----------|-----------|--------|
| MRCR 1M | 83.5% | Coming soon | Coming soon |
| CorpusQA 1M | 62.0% | Coming soon | Coming soon |
| ARC-AGI-1 | 90.00% | Coming soon | Coming soon |
| ARC-AGI-2 | 61.3% | Coming soon | Coming soon |
| MRCRv2 | Coming soon | 92.9% | Coming soon |
| LongBench v2 | Coming soon | 66.3% | Coming soon |

### Multimodal & Grounded

- Winner: Directional only
- DeepSeek V4 Pro public-lane score: 60.7 (Estimated · #54/119)
- Qwen3.8 Max public-lane score: 65.6 (Supported · #29/119)

| Benchmark | DeepSeek V4 Pro | 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: Qwen3.8 Max
- DeepSeek V4 Pro public-lane score: 59.7 (Supported · #46/158)
- Qwen3.8 Max public-lane score: 60.5 (Supported · #41/158)

| Benchmark | DeepSeek V4 Pro | Qwen3.8 Max | Winner |
|-----------|-----------|-----------|--------|
| MMLU-Pro | 87.5% | Coming soon | Coming soon |
| Chinese-SimpleQA | 84.4% | Coming soon | Coming soon |
| GPQA | 90.1% | 92.6% | Qwen3.8 Max |
| GPQA-D | 90.1% | 92.6% | Qwen3.8 Max |
| HLE | 42.7% | 43.6% | Qwen3.8 Max |
| GPQA Diamond (Vals) | 92.4% | 93.7% | Qwen3.8 Max |
| MMLU-Pro (Vals) | 87.0% | 88.6% | Qwen3.8 Max |
| HLE w/o tools | Coming soon | 43.6% | Coming soon |

### Multilingual

- Winner: Not comparable
- DeepSeek V4 Pro public-lane score: 57.3 (Estimated · #31/88)
- Qwen3.8 Max public-lane score: Coming soon

### Instruction Following

- Winner: Qwen3.8 Max
- DeepSeek V4 Pro public-lane score: 56.6 (Supported · #47/143)
- Qwen3.8 Max public-lane score: 60.9 (Supported · #33/143)

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

### Mathematics

- Winner: Qwen3.8 Max
- DeepSeek V4 Pro public-lane score: 59.4 (Supported · #34/125)
- Qwen3.8 Max public-lane score: 62.6 (Supported · #27/125)

| Benchmark | DeepSeek V4 Pro | Qwen3.8 Max | Winner |
|-----------|-----------|-----------|--------|
| HMMT Feb 2026 | 95.2% | Coming soon | Coming soon |
| IMOAnswerBench | 89.8% | Coming soon | Coming soon |
| Apex | 38.3% | Coming soon | Coming soon |
| Apex Shortlist | 90.2% | Coming soon | Coming soon |
| FrontierMath v2 (Tiers 1-3) | 64.561% | 74.737% | Qwen3.8 Max |
| FrontierMath v2 (Tier 4) | 26.829% | 46.341% | Qwen3.8 Max |

## FAQ

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

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

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

The widest category gap is in coding, where the averages are 59.6 for DeepSeek V4 Pro and 67.5 for Qwen3.8 Max.

### How many benchmarks does DeepSeek V4 Pro cover on BenchLM?

DeepSeek V4 Pro currently has 46 sourced benchmark scores on BenchLM.

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

Qwen3.8 Max currently has 70 sourced benchmark scores on BenchLM.

## Related Comparisons

- [DeepSeek V4 Pro vs Qwen3.8 Max Preview](/compare/deepseek-v4-pro-vs-qwen3-8-max-preview)
- [Qwen3.8 Max vs Qwen3.8 Max Preview](/compare/qwen3-8-max-vs-qwen3-8-max-preview)
- [DeepSeek V4 Pro vs Claude Fable 5.2](/compare/claude-fable-5-2-vs-deepseek-v4-pro)
- [Qwen3.8 Max vs Claude Fable 5.2](/compare/claude-fable-5-2-vs-qwen3-8-max)

## Explore More

- [DeepSeek V4 Pro profile](/models/deepseek-v4-pro)
- [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)
