Grok 4.1: benchmark results
xAI's Grok 4.1 model with improved conversational and emotional intelligence. Provider: xAI. Released 2025-11-01. Access: API.
Aggregate Rating 1688 ± 10, rank #227 of 1656 rated models, from 175 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CanMT | 5.18 | Mean translation score (1-7) over the 12 CanMT directions (u | 100 |
| CanMT - Chinese to Russian | 5.3 | Translation score (1-7) on the 95 Chinese-to-Russian CanMT s | 100 |
| CanMT - Chinese to Spanish | 5.25 | Translation score (1-7) on the 146 Chinese-to-Spanish CanMT | 100 |
| CanMT - English to Japanese | 5.34 | Translation score (1-7) on the 103 English-to-Japanese CanMT | 100 |
| CanMT - Russian to Chinese | 5 | Translation score (1-7) on the 125 Russian-to-Chinese CanMT | 100 |
| CanMT - Spanish to Chinese | 5.03 | Translation score (1-7) on the 166 Spanish-to-Chinese CanMT | 100 |
| SciAgentBench - L2 (4 to 7 Steps) | 43.9 | Success rate (%, 0-100) on the tasks of difficulty L2 (4 to | 100 |
| SciAgentBench - Physics | 47.2 | Success rate (%, 0-100) on the 109 physics tasks; with tools | 100 |
| SnitchBench | 92.5 | Govt Contact Rate (%) | 100 |
| URSA Single-Step Retrosynthesis | 0.49 | Average per-target ChemCensor score (0 to 5) over the model' | 100 |
| SciAgentBench - Materials Science | 32.4 | Success rate (%, 0-100) on the 37 materials science tasks; w | 96.9 |
| Chatbot Arena (Text - Medicine & Healthcare) | 1500 | Arena Score | 95.3 |
Interactive version: theaggregate.ai/model?slug=grok-4-1 · How It Works · Data refreshed daily, snapshot 2026-10-20.