DeepSeek-V4-Flash-0731 vs Qwen3.8 Max
How DeepSeek-V4-Flash-0731 and Qwen3.8 Max stack up across benchmarks, pricing, speed, and the workloads that matter.
DeepSeek-V4-Flash-0731
DeepSeek · Open weights
Qwen3.8 Max
Alibaba Cloud / Qwen Team · Proprietary
On agents & tool use, Qwen3.8 Max leads by about 25 points in our normalized benchmark average.
The table below breaks down every workload axis where both models have published results. Each score is a normalized 0 to 1 average across the benchmarks tagged with that category. A gap of a few points is noise; a gap of ten or more is a real difference in capability. Where one model has not reported a result, the cell shows n/a and the missing benchmark does not drag its average down.
Benchmark scores by workload
| Workload | DeepSeek-V4-Flash-0731 | Qwen3.8 Max | Edge |
|---|---|---|---|
| Reasoning | 51% | 74% | +22.9 pts |
| Coding | 66% | 68% | +2.1 pts |
| Agents & tool use | 54% | 79% | +25.4 pts |
| Math | n/a | 50% | n/a |
| Vision & multimodal | n/a | 80% | n/a |
| Long context | n/a | 74% | n/a |
| Writing | n/a | 75% | n/a |
| Knowledge & factuality | n/a | 82% | n/a |
The headline benchmarks below are the most widely cited individual tests. GPQA Diamond measures graduate-level reasoning in physics, chemistry, and biology. SWE-Bench Verified tests whether a model can fix real GitHub issues. MMMU-Pro covers college-level multimodal understanding across six disciplines. AIME and FrontierMath push competitive and research math. BrowseComp measures web research ability.
Headline benchmarks
| Benchmark | DeepSeek-V4-Flash-0731 | Qwen3.8 Max |
|---|---|---|
| GPQA | n/a | 93% |
| MMMU-Pro | n/a | 82% |
Benchmarks tell you what a model can do in a controlled setting. They do not tell you whether it will work on your specific task, with your specific data, at your specific scale. The recommendations below map each common workload to whichever of these two models scores higher on the relevant axis. Use them as a starting point, not a final answer.
Which should you choose?
Hard reasoning
Multi-step analysis, research, and problems that need sustained thought.
Qwen3.8 MaxCoding & software
Writing, reviewing, and debugging code across a real codebase.
Qwen3.8 MaxAgents & tool use
Long-running agents that call tools, browse, and act on their own.
Qwen3.8 MaxMath & science
Formal math, competitive problems, and quantitative science.
Qwen3.8 MaxVision & documents
Reading images, screenshots, charts, and dense documents.
Qwen3.8 MaxLong documents
Whole codebases, books, and transcripts that fill the context window.
Qwen3.8 MaxWriting & drafting
Copy, emails, and long-form drafting where tone matters.
Qwen3.8 MaxScores are normalized from public benchmarks published by llm-stats.com and averaged per workload. Pricing is the blended input/output cost per million tokens at an 8:1 mix. Refreshes daily. How this works.