tokenstat
homepricingstatshelpsign in
tokenstat

Compare

Kimi K3 vs Qwen3.8 Max

List rates and scores as of . Cheaper is lower $/MTok. Higher scores win. Input edge: Qwen3.8 Max Output edge: Qwen3.8 Max

Model A

Kimi K3

moonshotai/kimi-k3
reasoningtoolsopen
  • List input$2.70
  • List output$13.50
  • Context1.0M
  • Offers17
  • Arena code—
  • Arena agent9.91
  • Arena text—
  • Arena vision—
  • AA intelligence57.1
  • AA coding76.2
  • AA agentic50.1

On plans: OpenCode Go

Model B

Qwen3.8 Max

qwen/qwen3.8-max
reasoningtools
  • List input$2
  • List output$6
  • Context1M
  • Offers5
  • Arena code1668
  • Arena agent—
  • Arena text1496
  • Arena vision1305
  • AA intelligence—
  • AA coding—
  • AA agentic—

On plans: OpenCode Go

Score sources on /benchmarks.