DeepSeek V4 Flash 0420 (Max) vs Qwen3.8 Max (0902)
Which is better for the money: DeepSeek V4 Flash 0420 (Max) by DeepSeek or Qwen3.8 Max (0902) by Alibaba? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-10-24.
Trade-off
Verdict
Depends on your budget
Qwen3.8 Max (0902) scores 21.2 points higher (45.4 vs 24.2) but costs 33× as much to run the benchmark ($4935 vs $149; flagship vs mid-range cost classes). They are not alternatives at one budget: pick Qwen3.8 Max (0902) when you need intelligence above 24.2, and DeepSeek V4 Flash 0420 (Max) when 24.2 is enough.
| Metric | DeepSeek V4 Flash 0420 (Max) | Qwen3.8 Max (0902) | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 24.2 | 45.4 | +87.6% Qwen3.8 Max (0902) is 87.6% more intelligent |
| Total AA cost (lower is better) | $149.48 | $4,934.79 | +3201.3% Qwen3.8 Max (0902) is 33× more expensive |
| End-to-end time (lower is better) | 116.79s | 63.23s | −45.9% Qwen3.8 Max (0902) is 45.9% faster |
| Price per 1M tokens (in / out) | $0.11 / $0.23 | $2.00 / $6.00 | |
| Input modalities | Text | Text, Image |
Qwen3.8 Max (0902) vs DeepSeek V4 Flash 0420 (Max): 87.6% more intelligence, 3201.3% more cost to run the AA benchmark, and 45.9% faster end-to-end.
Today's Pareto frontier
Against the frontier
- DeepSeek V4 Flash 0420 (Max) is off the frontier: at the same or lower cost, GPT-6 Luna (Max) scores 38.10 for $122.
- Qwen3.8 Max (0902) is off the frontier: at the same or lower cost, Claude Opus 5.5 (Xhigh, Default Fallback) scores 56.00 for $4057.
DeepSeek V4 Flash 0420 (Max)Qwen3.8 Max (0902)Today's Pareto frontierOpen-weights frontierEvery other model
Matchup
Compare something else
Agents:
Accept: text/markdown or ?format=md returns this page as Markdown. JSON: /api/v1/compare/deepseek-v4-flash-0420/qwen3-8-max · Markdown: /api/v1/compare/deepseek-v4-flash-0420/qwen3-8-max?format=md · llms.txt · agent skill