Grok 4.3 (Non-reasoning): benchmark results
Provider: xAI. Released 2026-04-30. Access: API.
Aggregate Rating 1577 ± 16, rank #773 of 2168 rated models, from 93 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| FinBalance | 31 | Exact balance-sheet accuracy (%): share of records whose rep | 80 |
| Wolfram LLM Benchmarking Project | 53.4 | Correct Functionality (%) | 70.4 |
| Tau2-Bench Telecom | 65.79 | Success Rate (%) | 61.2 |
| AA TAU-2 Bench | 65.79 | Accuracy (%) | 59.9 |
| AA Terminal-Bench Hard | 18.94 | Accuracy (%) | 58.7 |
| AA-Omniscience Index - Software Engineering (SWE) - JavaScript | -11.82 | Omniscience Index | 58.1 |
| AA-Omniscience Hallucination Rate | 74.46 | Hallucination Rate (%) | 57.3 |
| AA-Omniscience Index - Law | -26.9 | Omniscience Index | 57.3 |
| AA-Omniscience Index - Software Engineering (SWE) - PHP | -18 | Omniscience Index | 56.9 |
| AA IFBench | 47.62 | Accuracy (%) | 56.3 |
| AA-Omniscience Index - Software Engineering (SWE) - R | -36 | Omniscience Index | 56 |
| AA-Omniscience Accuracy | 23.72 | Accuracy (%) | 54.9 |
Interactive version: theaggregate.ai/model?slug=grok-4-3-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-10-20.