ModelChorusModelChorus
ChallengeChatLeaderboardSwipeBenchmarksStore
OverviewVotersHead-to-HeadTextImageTTSSTTModels
How it worksAPITerms of ServicePrivacy Policy

Copyright 2026 MeetKai Inc.

Head-to-Head

Rankings tell you who’s on top — head-to-head tells you why. See how often any model beats another across all non-tied A vs. B challenges, and where the close calls are. Pick a model to view its complete record.

Average Win Rate Against All Other Models

Uniform sampling, excluding ties

ModelsBy avg win
100%
Focused model
MMiniMax M3
Average win rate across 2 opponents
Wins2100%
Losses00%
Ties0at 50%
Strongest vsGLM 5100%
Weakest vsDeepSeek V4 Pro100%
vs GLM 5100%
vs DeepSeek V4 Pro100%
vs Claude Haiku 4.5—
vs Claude Sonnet 4.5—
vs Claude Sonnet 4.6—
vs DeepSeek V4 Flash—
vs Functionary Swahili Large—
vs Functionary Swahili Mini—
vs GLM 5.2—
vs GPT-5 Nano—
vs GPT-5.2—
vs GPT-5.6 Luna—
vs GPT-oss-120B—
vs Kimi K2.6—
vs Mistral Medium 3.5—
vs Qwen3.7 Max—
vs Step 3.7 Flash—