Tools

Compare models

Add up to four models via ?models=slug-a,slug-b — or pick from any model page.

Spec
xAI
Context1M
License
Cheapest hosted (out/1M)$2.50 · xAI
ReleasedApr 30, 2026
LiveBench62.25
LiveBench Agentic Coding18.54
LiveBench Coding69.93
LiveBench Data Analysis55.77
LiveBench Instruction Following62.75
LiveBench Language73.58
LiveBench Mathematics84.34
LiveBench Reasoning70.82
Arena Agent54th of 55
Arena Agent · Recoverygets back on track after a command fails54th of 55
Arena Agent · Steerabilitydoes what it was asked, and changes course when told52nd of 55
Arena Agent · Task outcomefinishes what the session set out to do47th of 55
Arena Agent · Tool usereaches for the right one, and does not invent one7th of 55
Arena Coding1488
Arena Creative Writing1423
Arena Hard Prompts1456
Arena Instruction Following1415
Arena Maths1419
Arena Text (overall)1442
Arena Code (WebDev)1357