Qwen·released 2025-04-27✓ 2 sources
Qwen3-235B-A22B benchmark scores: 9 benchmarks tracked. 22% of its scores are independently reproduced — the rest are self-reported or unverified.
Leads on 0 of 9 benchmarks·best result 83 on Lech Mazur Writing (unverified)·2 of 9 independently reproduced·$0.2/$0.6 per M tokens
Consensus: LiteLLM · Cross-check: models.dev · See every model’s pricing →
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| Lech Mazur Writing | 83accuracy (%) | unverifiedT1 | 2025-04-28lechmazur/writing Github repository |
| MATH level 5 | 68.86accuracy (%) | reproduced· optimizedT1 | 2025-04-28Epoch AI |
| Fiction.LiveBench | 67.7accuracy (%) | unverifiedT1 | 2025-04-28Fiction.live leaderboard |
| GPQA diamond | 60.94accuracy (%) | reproduced· optimizedT1 | 2025-04-28Epoch AI |
| Aider polyglot | 59.6accuracy (%) | unverified· optimizedT1 | 2025-04-28Aider LLM Leaderboards |
| DTBench | 59.55accuracy (%) | unverifiedT1 | 2025-04-28https://conceptualreasoning.ai/dtbench |
| WeirdML | 37.28accuracy (%) | unverifiedT1 | 2025-04-28WeirdML Leaderboard |
| LMCA | 29.41accuracy (%) | unverifiedT1 | 2025-04-28https://conceptualreasoning.ai/lmca |
| SimpleBench | 17.2accuracy (%) | unverifiedT1 | 2025-04-28SimpleBench Leaderboard |
Claims drawn from cited facts, not live model generation.
Developed by Qwen, this model is a large-scale model with 235.09 billion parameters, released in 2025. That scale implies greater headroom and higher expense, whereas smaller models would offer efficiency and lower cost.
4 cited facts
This model is scored on 9 tracked benchmarks and holds no current top score. Its record has been independently reproduced, so the scores are backed by outside evaluation.
3 cited facts
On average it trails the best reported score by 35.9 points under the disclosed evaluation harness, a wide gap that reads as well behind the front of the pack rather than a near-frontier result. It holds the top score on 0 benchmarks, so there is no category where it currently leads.
2 cited facts
6 facts cross-checked across data sources: 1 corroborated, 1 single-source, 4 disputed.
Source facts, citations, and refresh stamp for this record.
Sources: epoch_benchmarks, huggingface_models, litellm_prices, modelsdev_models
Qwen3-235B-A22B is an AI model developed by Qwen, released 2025-04-27. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.
Qwen3-235B-A22B has recorded scores on 9 benchmarks, each shown with its evidence status.
Qwen3-235B-A22B has recorded scores on 9 benchmarks — Lech Mazur Writing, GPQA diamond, MATH level 5, DTBench, WeirdML, LMCA, and 3 more. The full table above shows each score with its evidence status.
2 of 9 recorded scores (22%) are independently reproduced rather than self-reported by the lab.
Qwen3-235B-A22B has 9 tracked claims: 2 independently reproduced, 7 unverified.
Listed API pricing: $0.2 per million input tokens, $0.6 per million output tokens (prices disputed across sources). See the pricing block for the full breakdown.