kimi-k3 vs claude-sonnet-5
Pricing, Performance & Features Comparison
Moonshot’s 2.8T-parameter flagship for long-horizon coding and knowledge work, with native vision and a 1M-token context window; thinking is always on.
Input$3
Output$15
Latency (p50)5.8s
Output Limit1M
Function Calling
JSON Mode
InputText, Image, Video
OutputText
in$3out$15cache$0.3–
Latency (24h)
Success Rate (24h)
Near-Opus quality on coding and agentic work at Sonnet cost; adaptive thinking is on by default and non-default sampling parameters are rejected.
Input$3
Output$15
Latency (p50)5.6s
Output Limit128K
Function Calling
JSON Mode
InputText, Image
OutputText
in$3out$15cache$0.3write$3.8