Grok 3
Sample reasoning profile for comparison workflows.
Source and version
This page uses phase-1 sample data. Before launch, every data point must show official or measured source, crawl/test time, benchmark version and methodology link.
- Source type: sample-estimated
- Benchmark version: api-2026-06-sample
- Last updated: 2026-06-09 22:40 CST
Task performance
Repository edits, bug fixes, unit-test reasoning and API usage.
72WritingProduct copy, long-form prose, editing and tone control.
68TranslationBidirectional English/Chinese translation with terminology consistency.
69MathWord problems, algebraic reasoning and structured calculation.
78Tool CallingJSON output, function selection and multi-step tool use.
72ChineseChinese writing, knowledge, instructions and domestic model coverage.
68Use and disclosure
The use link is a placeholder. Rankings must never be affected by affiliate relationships, and production affiliate links require clear disclosure.
Price alert
Similar models
- No sibling models yet.