Mistral Large 4 Preview vs Mistral Small 3.2
Which is better for the money: Mistral Large 4 Preview by Mistral or Mistral Small 3.2 by Mistral? 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
Mistral Large 4 Preview scores 31.4 points higher (38.4 vs 7.0) but costs 13× as much to run the benchmark ($1608 vs $122; premium vs mid-range cost classes). They are not alternatives at one budget: pick Mistral Large 4 Preview when you need intelligence above 7.0, and Mistral Small 3.2 when 7.0 is enough.
| Metric | Mistral Large 4 Preview | Mistral Small 3.2 | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 38.4 | 7.0 | −81.8% Mistral Small 3.2 is 81.8% less intelligent |
| Total AA cost (lower is better) | $1,607.86 | $121.78 | −92.4% Mistral Small 3.2 is 92.4% cheaper |
| End-to-end time (lower is better) | 27.12s | 12.50s | −53.9% Mistral Small 3.2 is 53.9% faster |
| Price per 1M tokens (in / out) | $1.36 / $4.18 | $0.07 / $0.20 | |
| Input modalities | Text, Image | Text, Image |
Mistral Small 3.2 vs Mistral Large 4 Preview: 81.8% less intelligence, 92.4% less cost to run the AA benchmark, and 53.9% faster end-to-end.
Today's Pareto frontier
Against the frontier
- Mistral Large 4 Preview is off the frontier: at the same or lower cost, GPT-6.1 Sol (Max) scores 51.80 for $1082.
- Mistral Small 3.2 is off the frontier: at the same or lower cost, MiMo-V2.6-Flash scores 37.90 for $110.
Mistral Large 4 PreviewMistral Small 3.2Today'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/mistral-large-4/mistral-small-3-2 · Markdown: /api/v1/compare/mistral-large-4/mistral-small-3-2?format=md · llms.txt · agent skill