Back
Kimi

Kimi K2.6

MoonshotAI, China185/191 correctLast updated May 20, 2026Website

Kimi K2.6 by MoonshotAI achieved an overall score of 96.3% on Appwrite Arena, ranking #4 of 12 benchmarked models. The benchmark tests how well AI models understand Appwrite - the open-source backend platform for authentication, TablesDB, storage, functions, and more. This model answered 185 of 191 questions correctly across categories including Auth, TablesDB, Functions, Storage, and CLI. During the benchmark it consumed 1,687,080 tokens costing $1.6431, averaging 17.0 tokens per second over 134m 48s. Compare Kimi K2.6 with other LLMs on the Appwrite Arena leaderboard.

Total tokens
1,687,080Input: 1,549,720 Output: 137,360
Total duration
134m 48s
Avg speed
17.0 tok/s
Total cost
$ 1.6431
Cost/1MOverallFoundationAuthTablesDBFunctionsStorageSitesMessagingRealtimeCLI
$0.73Input: $0.73/1M tokens
Output: $3.49/1M tokens		96.3%99.7%99%90.9%94.5%90%98.5%99.5%99%96.3%
Deterministic17/17 correct
AI-Judged98% average score