HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks
Home/Orgs/Qwen/Qwen/Qwen3-4B-Thinking-2507-FP8

Qwen/Qwen3-4B-Thinking-2507-FP8

family qwen · 1 variants · unknown

provider
Qwen
released
—
huggingface
https://huggingface.co/Qwen/Qwen3-4B-Thinking-2507-FP8
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 (1)

variantkindquantbitsvRAMdlorgyear
Qwen3-4B-Thinking-2507-FP8fp8fp88.0178unsloth2025
Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.