DeepSeek V3 (Dec '24) vs Ministral 3 3B
Which is better for the money: DeepSeek V3 (Dec '24) by DeepSeek or Ministral 3 3B by Mistral? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-10-25.
Trade-off
Verdict
Depends on your budget
DeepSeek V3 (Dec '24) scores 3.7 points higher (8.5 vs 4.8) but costs 2.1× as much to run the benchmark ($25.4 vs $12.2; both budget class). Pick DeepSeek V3 (Dec '24) when the extra intelligence is worth the premium, otherwise Ministral 3 3B.
| Metric | DeepSeek V3 (Dec '24) | Ministral 3 3B | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 8.5 | 4.8 | −43.5% Ministral 3 3B is 43.5% less intelligent |
| Total AA cost (lower is better) | $25.37 | $12.23 | −51.8% Ministral 3 3B is 51.8% cheaper |
| End-to-end time (lower is better) | 13.85s | 3.01s | −78.3% Ministral 3 3B is 78.3% faster |
| Price per 1M tokens (in / out) | $0.32 / $0.89 | $0.10 / $0.10 | |
| Input modalities | Text | Text, Image |
Ministral 3 3B vs DeepSeek V3 (Dec '24): 43.5% less intelligence, 51.8% less cost to run the AA benchmark, and 78.3% faster end-to-end.
Today's Pareto frontier
Against the frontier
- DeepSeek V3 (Dec '24) is off the frontier: at the same or lower cost, GPT-6 Luna (Low) scores 21.50 for $10.6.
- Ministral 3 3B is off the frontier: at the same or lower cost, GPT-6 Luna (Low) scores 21.50 for $10.6.
DeepSeek V3 (Dec '24)Ministral 3 3BToday'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-v3/ministral-3-3b · Markdown: /api/v1/compare/deepseek-v3/ministral-3-3b?format=md · llms.txt · agent skill