Navigate the site or open source-checked model and benchmark registry records.
Navigate
Lumina Bench
News
Research
Independent rankings. Commercial relationships never affect scoring.
xAI · stable
Grok 4 is tracked via Artificial Analysis independent evaluations for capability comparison across coding, agents, reasoning and related workloads.
#51 · 61.9 score · Confidence B