Qwen 3 235B A22B FP8 — benchmark results
Qwen 3 235B A22B evaluated in FP8 quantization. Provider: Alibaba. Released 2025-04-28. Access: Open.
Unified ELO 1606 ± 29, rank #468 of 1776 rated models, from 14 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Kluster Hallucination Detection - RAG Method 3 Resistance | 93.37 | Resistance (100 - Hallucination Rate %) | 70 |
| ReasonScape M12X | 473.79 | ReasonScore (hard) | 65.8 |
| Kluster Hallucination Detection - RAG Method 1 Resistance | 93.2 | Resistance (100 - Hallucination Rate %) | 63.6 |
| StickToYourRole | 71.84 | Cardinal Score | 62.9 |
| BenchTable | 55.5 | Total Score (%) | 62.8 |
| LMGame-Bench Candy Crush | 363.3 | Score | 50 |
| LMGame-Bench 2048 | 2144 | Score | 45.8 |
| LMGame-Bench Sokoban | 2.3 | Score | 45.8 |
| Kluster Hallucination Detection - Non-RAG Resistance | 97.91 | Resistance (100 - Hallucination Rate %) | 43.8 |
| LLM Hallucination Detection - Non-RAG | 2.09 | Hallucination rate (%) | 43.8 |
| Kluster Hallucination Detection - RAG Method 2 Resistance | 94.96 | Resistance (100 - Hallucination Rate %) | 36.4 |
| Kluster Hallucination Detection - RAG Resistance | 93.84 | Resistance (100 - Hallucination Rate %) | 31.2 |
Interactive version: theaggregate.ai/model?slug=qwen-3-235b-a22b-fp8 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.