HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks
Home/Orgs/Qwen/Qwen/Qwen3-235B-A22B-Instruct-2507

Qwen/Qwen3-235B-A22B-Instruct-2507

family qwen · 9 variants · unknown

provider
Qwen
released
—
huggingface
https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507
best scores
MMLU-Pro: 81.2MMLU: 90.4GPQA: 68.4SimpleQA: 71.1AIME 2025: 46.6ARC-AGI: 9.0LiveCodeBench: 45.2IFEval: 82.3Chatbot Arena: 45.6BFCL (function calling): 64.7

Best scores per benchmark

benchmarkcategorybest score
MMLUknowledge90.4
IFEvalinstruction-following82.3
MMLU-Proknowledge81.2
SimpleQAknowledge71.1
GPQAreasoning68.4
BFCL (function calling)agentic64.7
AIME 2025math46.6
Chatbot Arenachat45.6
LiveCodeBenchcoding45.2
ARC-AGIreasoning9.0

Variants (9)

variantkindquantbitsvRAMdlorgyear
Qwen3-235B-A22B-Instruct-2507-FP8fp8fp88.086,700Qwen2025
Qwen3-235B-A22B-Instruct-2507originalbf1616.064,434Qwen2025
Qwen3-235B-A22B-Instruct-2507-NVFP4nvfp4nvfp44.014,506nvidia2026
Qwen3-235B-A22B-Instruct-2507-GGUFggufgguf16.09,588unsloth2025
Qwen3-235B-A22B-Instruct-2507-MLX-4bitmlxmlx3,496lmstudio-community2025
Qwen3-235B-A22B-Instruct-2507-MLX-8bitmlxmlx3,261lmstudio-community2025
Qwen3-235B-A22B-Instruct-2507-MLX-6bitmlxmlx3,232lmstudio-community2025
Qwen3-235B-A22B-Instruct-2507-GGUFggufgguf3.02,539lmstudio-community2025
Qwen_Qwen3-235B-A22B-Instruct-2507-GGUFggufgguf1.01,990bartowski2025
Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.