Kimi K2
OldMoonshot
•
Open Source•
1T# 54
Released
Jul 11, 2025
# 11
Context Length
131K
Benchmarks
# 98
Code RankedAGI
48.3%
# 30
SWEBench Verified
71.6%
# 125
Agentic RankedAGI
38.8%
# 35
LiveCodeBench v6
53.7%
# 12
Coding LiveBench 25.5
71.8%
# 19
Aider Polyglot
60.0%
# 134
Reason RankedAGI
41.5%
# 75
HLE
4.7%
# 58
GPQA Diamond
75.1%
# 14
Reason LiveBench 25.5
63.0%
# 54
AIME 2025 I & II
49.5%
# 26
AIME 2024
69.6%
# 16
Math LiveBench 25.5
74.4%
# 20
MMLU Pro
81.1%
# 9
MMLU
89.5%
# 5
IF LiveBench 25.5
82.5%
# 15
Avg LiveBench 25.5
62.7%
# 5
IF Evaluation
89.8%
# 9
Agentic LiveBench 25.5
20.0%
# 125
Math RankedAGI
54.5%
# 123
RAGI RankedAGI
45.8%
# 67
GDPval AA
529
# 3
Svelte Bench v1
84.4%
# 46
Code DesignArena
1098
Pricing
# 25
Input Cost /M
$0.6
# 30
Output Cost /M
$2.5
# 16
Cached Cost /M
$0.15