MiMo-V2.5 vs Qwen3.8 Max
Which is better for the money: MiMo-V2.5 by Xiaomi or Qwen3.8 Max 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 scores 18.4 points higher (40.2 vs 21.8) but costs 51× as much to run the benchmark ($3172 vs $61.9; flagship vs mid-range cost classes). They are not alternatives at one budget: pick Qwen3.8 Max when you need intelligence above 21.8, and MiMo-V2.5 when 21.8 is enough.
| Metric | MiMo-V2.5 | Qwen3.8 Max | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 21.8 | 40.2 | +84.4% Qwen3.8 Max is 84.4% more intelligent |
| Total AA cost (lower is better) | $61.93 | $3,172.37 | +5022.5% Qwen3.8 Max is 51× more expensive |
| End-to-end time (lower is better) | 41.73s | 62.83s | +50.6% Qwen3.8 Max is 50.6% slower |
| Price per 1M tokens (in / out) | $0.13 / $0.27 | $2.00 / $6.00 | |
| Input modalities | Text, Image | Text, Image, Video |
Qwen3.8 Max vs MiMo-V2.5: 84.4% more intelligence, 5022.5% more cost to run the AA benchmark, and 50.6% slower end-to-end.
Today's Pareto frontier
Against the frontier
- MiMo-V2.5 is off the frontier: at the same or lower cost, Claude Haiku 5.5 (Medium) scores 34.50 for $56.4.
- Qwen3.8 Max is off the frontier: at the same or lower cost, Kimi K3.1 Pro scores 55.10 for $3117.
MiMo-V2.5Qwen3.8 MaxToday'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/mimo-v2-5-0424/qwen3-8-max-0803 · Markdown: /api/v1/compare/mimo-v2-5-0424/qwen3-8-max-0803?format=md · llms.txt · agent skill