Grok Beta: benchmark results
xAI's grok-beta, the Grok-2-class 128K-context preview model that launched the public xAI API beta in November 2024. Provider: xAI. Released 2024-11-04. Access: API.
Unified ELO 1534 ± 1, rank #519 of 1392 rated models, from 29 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SpeechMap Compliance | 93.7 | % Requests Completed | 96.2 |
| TextClass Benchmark | 1741.94 | Meta-Elo (self-reported) | 91.6 |
| ProLLM - Function Calling | 91.8 | Score (%) | 90.7 |
| VNTL Leaderboard | 71.27 | Accuracy (%) | 89.5 |
| EvalPlus (HumanEval+ & MBPP+) | 73 | Pass@1 avg (%) | 86.3 |
| ProLLM - Summarization | 74.1 | Score (%) | 61.2 |
| ForecastBench | 64.3 | Overall Score (higher is better) | 61 |
| AI for Education Pedagogy - Technology | 80.19 | Accuracy (%) | 57.4 |
| HumanEval+ | 80.5 | HumanEval+ pass@1 (self-reported) | 56.5 |
| ProLLM - StackEval | 90.4 | Score (%) | 51.2 |
| NYT Connections Original | 23.7 | Score (%) | 48.3 |
| ProLLM - Q&A Assistant | 93.1 | Score (%) | 44.1 |
Interactive version: theaggregate.ai/model?slug=grok-beta · How It Works · Data refreshed daily, snapshot 2026-09-05.