Mo
kimi-k2.6
Benchmark & Insights
Moonshot AI Cloudflare Workers AI
Updated Oct 4, 2026 All models
Sample size
232 runs
in window
Accuracy
70.5%
reference match · 241d
Confidence
90%
over 242 runs
Window end
Oct 4, 2026
most recent run
Input price
$0.95/MTok
prompt tokens
Output price
$4.00/MTok
completion tokens
Model insights
- 01 Much weaker than kimi-k2.5 suggested: over the full period it leans low, calling 40 "MEDIUM_SAFE" days "low".
- 02 It called 6 of 19 "MEDIUM_UNSAFE" days "low" and missed 10 "unsafe" days.
- 03 Pricier than k2.5 at $0.95/$4 with no gain.
Notes
Grandfathering a silver tier leader Kimi K2.5