Skip to main content

MoonshotAI: Kimi K2.6

MoonshotAI: Kimi K2.6 is moonshotai's model. On the un-leaked meo benchmark it ranks #9 of 22 with 54.6% accuracy and an Effective Value (๐•) of 0.0002 โ€” earning a correct answer for about $0.0440 in 251s each.

as of 2026-06-08

meo rank
#9
of 22
Accuracy
54.6%
overall
๐• index (top=100)
0.32
raw ๐• 0.0002 ยท N=10
$ / correct
$0.0440
251s each

Per-domain accuracy

CSP (unsatisfiable)
100%
Tape machine
100%
Theory of mind
100%
Regex automaton
88%
Framework bias
67%
Perceptual illusions
51%
Long arithmetic
33%
Watson-Glaser
25%
Logic / math / CS
8%
Base-rate bias
2%
State tracking
0%

Efficiency

Cost per correct
$0.0440
Seconds per correct
251
Tokens per correct
11,698
Correct answers
111 / 251
Total benchmark cost
$4.88
Avg tokens / item
5,173

Model details

Lab
moonshotai
Released
2026-04-20
Weights
Open
Context
262K
Max output
262K
Knowledge cutoff
โ€”
Price in
$0.68/M
Price out
$3.42/M
Modality
text+image->text

Sources: openrouter ยท meo

FAQ

How accurate is MoonshotAI: Kimi K2.6?

MoonshotAI: Kimi K2.6 scores 54.6% overall on the meo benchmark (rank #9 of 22), an un-leaked private holdout graded objective-first.

What is MoonshotAI: Kimi K2.6's Effective Value?

Its Effective Value (๐•) is 0.0002 at N=10, ฯ‰=1, ฮด=1.5 โ€” a fused metric of accuracy, speed, cost, and error-cascade. ๐• rankings shift with task depth.

How much does MoonshotAI: Kimi K2.6 cost per correct answer?

About $0.0440 per correct answer (11,698 tokens, 251 seconds each) at benchmark-time pricing via OpenRouter.

Is MoonshotAI: Kimi K2.6 open weights?

Yes โ€” MoonshotAI: Kimi K2.6 is released with open weights.