Granite 4.2 3B vs Kimi K3 (Low)
Which is better for the money: Granite 4.2 3B by IBM or Kimi K3 (Low) by Kimi? 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
Kimi K3 (Low) scores 21.0 points higher (30.1 vs 9.1) but costs 80× as much to run the benchmark ($1190 vs $14.9; premium vs budget cost classes). They are not alternatives at one budget: pick Kimi K3 (Low) when you need intelligence above 9.1, and Granite 4.2 3B when 9.1 is enough.
| Metric | Granite 4.2 3B | Kimi K3 (Low) | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 9.1 | 30.1 | +230.8% Kimi K3 (Low) is 3.3× more intelligent |
| Total AA cost (lower is better) | $14.91 | $1,190.42 | +7884.0% Kimi K3 (Low) is 80× more expensive |
| End-to-end time (lower is better) | 35.30s | 66.33s | +87.9% Kimi K3 (Low) is 87.9% slower |
| Price per 1M tokens (in / out) | $0.03 / $0.12 | $3.00 / $15.0 | |
| Input modalities | Text | Text, Image |
Kimi K3 (Low) vs Granite 4.2 3B: 230.8% more intelligence, 7884.0% more cost to run the AA benchmark, and 87.9% 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.
- Kimi K3 (Low) is off the frontier: at the same or lower cost, GPT-6.1 Sol (Max) scores 51.80 for $1082.
Granite 4.2 3BKimi K3 (Low)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/granite-4-2-3b/kimi-k3-low · Markdown: /api/v1/compare/granite-4-2-3b/kimi-k3-low?format=md · llms.txt · agent skill