Grok 4
von xAI · Veroeffentlicht 2025-07-09
54.8
Durchschn. Score
$3.00/1M
Eingabepreis
$15.00/1M
Ausgabepreis
256K tokens (~128 books)
Kontextfenster
multimodal
Typ
Tested on 24 benchmarks with 54.8% average. Top scores: HELM — IFEval (94.9%), Fiction.LiveBench (94.4%), HELM — MMLU-Pro (85.1%).
Benchmark-Ergebnisse
| Benchmark | Kategorie | Score | Bar |
|---|---|---|---|
| HELM — IFEval | language | 94.9 | |
| Fiction.LiveBench | knowledge | 94.4 | |
| HELM — MMLU-Pro | knowledge | 85.1 | |
| OTIS Mock AIME 2024-2025 | math | 84.0 | |
| GPQA diamond | knowledge | 82.7 | |
| Lech Mazur Writing | knowledge | 80.7 | |
| HELM — WildBench | reasoning | 79.7 | |
| Aider polyglot | coding | 79.6 | |
| HELM — GPQA | knowledge | 72.6 | |
| ARC-AGI | reasoning | 66.7 | |
| HELM — Omni-MATH | math | 60.3 | |
| SimpleBench | reasoning | 52.6 | |
| DeepResearch Bench | knowledge | 47.9 | |
| SimpleQA Verified | knowledge | 47.9 | |
| WeirdML | coding | 45.7 | |
| GeoBench | knowledge | 45.0 | |
| Balrog | knowledge | 43.6 | |
| Cybench | coding | 43.0 | |
| Chess Puzzles | knowledge | 28.0 | |
| Terminal Bench | coding | 27.2 | |
| FrontierMath-2025-02-28-Private | math | 19.7 | |
| ARC-AGI-2 | reasoning | 16.0 | |
| APEX-Agents | agentic | 15.2 | |
| FrontierMath-Tier-4-2025-07-01-Private | math | 2.1 |