HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks
Home/Orgs/Qwen/Qwen/Qwen3-VL-30B-A3B-Thinking

Qwen/Qwen3-VL-30B-A3B-Thinking

family qwen · 12 variants · unknown

provider
Qwen
released
—
huggingface
https://huggingface.co/Qwen/Qwen3-VL-30B-A3B-Thinking
best scores
WebArena: 46.3OSWorld: 76.1

Best scores per benchmark

benchmarkcategorybest score
OSWorldagentic76.1
WebArenaagentic46.3

Variants (12)

variantkindquantbitsvRAMdlorgyear
Qwen3-VL-30B-A3B-Thinking-FP8fp8fp88.043,718Qwen2025
Qwen3-VL-30B-A3B-Thinkingoriginalbf1616.016,474Qwen2025
Qwen3-VL-30B-A3B-Thinking-GGUFggufgguf16.04,070unsloth2026
Qwen_Qwen3-VL-30B-A3B-Thinking-GGUFggufgguf2.02,665bartowski2025
Qwen3-VL-30B-A3B-Thinking-GGUFggufgguf16.02,533Qwen2025
Qwen3-VL-30B-A3B-Thinking-GGUFggufgguf4.02,150lmstudio-community2025
Qwen3-VL-30B-A3B-Thinking-1M-GGUFggufgguf16.01,306unsloth2025
Qwen3-VL-30B-A3B-Thinking-MLX-8bitmlxmlx305lmstudio-community2025
Qwen3-VL-30B-A3B-Thinking-MLX-4bitmlxmlx256lmstudio-community2025
Holo2-30B-A3B-GGUFggufgguf16.0220Mungert2025
Qwen3-VL-30B-A3B-Thinking-MLX-6bitmlxmlx169lmstudio-community2025
Qwen3-VL-30B-A3B-Thinking-MLX-5bitmlxmlx104lmstudio-community2025
Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.