xAI
Grok 4 benchmark results
13 results on 7 benchmarks from 4 sources.
Closed Released: Jul 9, 2025 Updated Oct 22, 2026
Also listed as: grok-4-0709 · Grok-4-0709 (Prompt) · Grok-4-0709 (FC) · Grok-4
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1411 | 147 | Overall, style control on; 39691 votes; 95% CI 1407–1415“grok-4-0709” | Official leaderboard | Oct 8, 2026 |
| LMArena Vision | 1184 | 96 | Overall, style control on; 30184 votes; 95% CI 1177–1192“grok-4-0709” | Official leaderboard | Oct 2, 2026 |
| GPQA Diamond | 87.0% | 72 | Epoch AI run, accuracy | Independent run | Jul 9, 2025 |
| MMLU-Pro | 87.0% | 18 | Overall accuracy; data source: Self-Reported“Grok-4” | Vendor-reported | Mar 11, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 84.0% | 101 | Epoch AI run, OTIS Mock AIME 2024–2025 | Independent run | Jul 9, 2025 |
| Terminal-Bench 2.0 | 27.2% | 140 | Terminal-Bench 2.0; agent: OpenHands | Official leaderboard via Epoch AI | Nov 2, 2025 |
| Terminal-Bench 2.0 | 27.2% | 141 | Terminal-Bench 2.0; agent: OpenHands | Official leaderboard via Epoch AI | Jul 9, 2025 |
| Terminal-Bench 2.0 | 25.4% | 149 | Terminal-Bench 2.0; agent: Mini-SWE-Agent | Official leaderboard via Epoch AI | Nov 3, 2025 |
| Terminal-Bench 2.0 | 25.4% | 151 | Terminal-Bench 2.0; agent: Mini-SWE-Agent | Official leaderboard via Epoch AI | Jul 9, 2025 |
| Terminal-Bench 2.0 | 23.1% | 160 | Terminal-Bench 2.0; agent: Terminus 2 | Official leaderboard via Epoch AI | Jul 9, 2025 |
| Terminal-Bench 2.0 | 23.1% | 161 | Terminal-Bench 2.0; agent: Terminus 2 | Official leaderboard via Epoch AI | Oct 31, 2025 |
| BFCL v4 (function calling) | 63.0% | 9 | Overall accuracy; prompt-based“Grok-4-0709 (Prompt)” | Official leaderboard | Oct 17, 2026 |
| BFCL v4 (function calling) | 61.4% | 10 | Overall accuracy; native function calling (FC)“Grok-4-0709 (FC)” | Official leaderboard | Oct 17, 2026 |
FAQ
Where do these Grok 4 scores come from?
From Epoch AI Benchmarking Hub, Berkeley Function Calling Leaderboard (BFCL), LMArena leaderboard dataset, and MMLU-Pro leaderboard (TIGER-Lab). Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.