Granite 4.2 3B vs MiMo-V2.5
Which is better for the money: Granite 4.2 3B by IBM or MiMo-V2.5 by Xiaomi? 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
MiMo-V2.5 scores 12.7 points higher (21.8 vs 9.1) but costs 4.2× as much to run the benchmark ($61.9 vs $14.9; mid-range vs budget cost classes). They are not alternatives at one budget: pick MiMo-V2.5 when you need intelligence above 9.1, and Granite 4.2 3B when 9.1 is enough.
| Metric | Granite 4.2 3B | MiMo-V2.5 | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 9.1 | 21.8 | +139.6% MiMo-V2.5 is 2.4× more intelligent |
| Total AA cost (lower is better) | $14.91 | $61.93 | +315.4% MiMo-V2.5 is 4.2× more expensive |
| End-to-end time (lower is better) | 35.30s | 41.73s | +18.2% MiMo-V2.5 is 18.2% slower |
| Price per 1M tokens (in / out) | $0.03 / $0.12 | $0.13 / $0.27 | |
| Input modalities | Text | Text, Image |
MiMo-V2.5 vs Granite 4.2 3B: 139.6% more intelligence, 315.4% more cost to run the AA benchmark, and 18.2% slower end-to-end.
Today's Pareto frontier
Against the frontier
- Granite 4.2 3B is off the frontier: at the same or lower cost, GPT-6 Luna (Low) scores 21.50 for $10.6.
- MiMo-V2.5 is off the frontier: at the same or lower cost, Claude Haiku 5.5 (Medium) scores 34.50 for $56.4.
Granite 4.2 3BMiMo-V2.5Today'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/granite-4-2-3b/mimo-v2-5-0424 · Markdown: /api/v1/compare/granite-4-2-3b/mimo-v2-5-0424?format=md · llms.txt · agent skill