xAI's reasoning-native flagship model, which uses tools during its chain of thought by default. xAI publishes its benchmark results only as chart images, so no sourced scores are recorded here yet.

Benchmarks

What its maker published, and what anyone else measured. Every figure links the document it came from.

benchmark	published	measured
MMLU-Pro	—	86.6
GPQA	—	87.7
AIME 2025	—	92.7
HLE	—	23.9

Lineage

↓ finetune Grok 4.1

view in full graph →

Metadata

org	xAI
released	2025-07-09
license	proprietary