HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks
Home/Orgs/Qwen/Qwen/Qwen3-4B-Thinking-2507

Qwen/Qwen3-4B-Thinking-2507

family qwen · 16 variants · unknown

provider
Qwen
released
—
huggingface
https://huggingface.co/Qwen/Qwen3-4B-Thinking-2507
best scores
MMLU-Pro: 78.5MMLU: 89.5GPQA: 65.8AIME 2025: 70.9LiveCodeBench: 57.4IFEval: 86.5Chatbot Arena: 36.3BFCL (function calling): 69.1

Best scores per benchmark

benchmarkcategorybest score
MMLUknowledge89.5
IFEvalinstruction-following86.5
MMLU-Proknowledge78.5
AIME 2025math70.9
BFCL (function calling)agentic69.1
GPQAreasoning65.8
LiveCodeBenchcoding57.4
Chatbot Arenachat36.3

Variants (16)

variantkindquantbitsvRAMdlorgyear
Qwen3-4B-Thinking-2507originalbf1616.0294,867Qwen2025
Qwen3-4B-Thinking-2507-FP8fp8fp88.0206,936Qwen2025
Qwen3-4B-Thinking-2507-MLX-4bitmlxmlx51,394lmstudio-community2025
Qwen3-4B-Thinking-2507-MLX-8bitmlxmlx50,643lmstudio-community2025
Qwen3-4B-Thinking-2507-MLX-6bitmlxmlx50,306lmstudio-community2025
Qwen3-4B-Thinking-2507-GGUFggufgguf3.015,305lmstudio-community2025
Qwen3-4B-Thinking-2507-unsloth-bnb-4bitbnb-4bit4bit4,951unsloth2025
Qwen3-4B-Thinking-2507-GGUFggufgguf16.04,738unsloth2025
Qwen_Qwen3-4B-Thinking-2507-GGUFggufgguf2.03,056bartowski2025
Qwen3-4B-Thinking-2507originalbf1616.02,498unsloth2025
Qwen3-4B-Thinking-2507-bnb-4bitbnb-4bit4bit641unsloth2025
DMind-3-mini-GGUFggufgguf16.0429Mungert2026
Jan-v1-4B-GGUFggufgguf16.0180Mungert2025
Qwen3-4B-Thinking-2507-Q8_0-GGUFggufgguf8.0136ggml-org2025
Qwen3-4B-Thinking-2507-MLX-5bitmlxmlx72lmstudio-community2025
Qwen3-4B-Thinking-2507-GGUFggufgguf16.017Mungert2025
Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.