Kimi K2.7 Code vs Quasar 438B
Which is better for the money: Kimi K2.7 Code by Kimi or Quasar 438B by Multiverse Computing? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-09-01.
Clear winner
Verdict
Pick Kimi K2.7 Code
Quasar 438B vs Kimi K2.7 Code: the same intelligence, 92.4% more cost to run the AA benchmark, and 75.1% faster end-to-end.
| Metric | Kimi K2.7 Code | Quasar 438B | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 43.0 | 43.0 | Even |
| Total AA cost (lower is better) | $544.58 | $1,047.71 | +92.4% Quasar 438B is 92.4% more expensive |
| End-to-end time (lower is better) | 60.38s | 15.01s | −75.1% Quasar 438B is 75.1% faster |
| Price per 1M tokens (in / out) | $0.95 / $4.00 | $0.60 / $1.80 | |
| Input modalities | Not reported | Not reported |
Today's Pareto frontier
Against the frontier
- Kimi K2.7 Code is off the frontier: at the same or lower cost, GLM-5.3-Flash scores 57.50 for $138.
- Quasar 438B is off the frontier: at the same or lower cost, Grok 4.6 (medium) scores 59.00 for $930.
Kimi K2.7 CodeQuasar 438BToday'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/kimi-k2-7-code/quasar-438b · Markdown: /api/v1/compare/kimi-k2-7-code/quasar-438b?format=md · llms.txt · agent skill