HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks
Home/Orgs/Qwen/Qwen/Qwen2.5-VL-7B-Instruct

Qwen/Qwen2.5-VL-7B-Instruct

family qwen · 32 variants · unknown

provider
Qwen
released
—
huggingface
https://huggingface.co/Qwen/Qwen2.5-VL-7B-Instruct
best scores
MMMU: 56.0DocVQA: 96.0ChartQA: 84.8MMBench: 87.9MT-Bench: 63.7MathVista: 73.8

Best scores per benchmark

benchmarkcategorybest score
DocVQAmultimodal96.0
MMBenchmultimodal87.9
ChartQAmultimodal84.8
MathVistamultimodal73.8
MT-Benchchat63.7
MMMUmultimodal56.0

Variants (32)

variantkindquantbitsvRAMdlorgyear
Qwen2.5-VL-7B-Instructoriginalbf1616.09,148,348Qwen2025
Qwen2.5-VL-7B-Instruct-AWQawqawq2.52,748,113Qwen2025
olmOCR-2-7B-1025-FP8fp8fp88.0699,201allenai2026
Qwen2.5-VL-7B-Instruct-GGUFggufgguf16.0188,724unsloth2025
Qwen2.5-VL-7B-Instruct-GGUFggufgguf3.069,084lmstudio-community2025
olmOCR-2-7B-1025originalbf1616.043,616allenai2025
Cosmos-Reason1-7Boriginalbf1616.028,385nvidia2025
Qwen2.5-VL-7B-Instruct-NVFP4nvfp4nvfp44.025,660nvidia2025
olmOCR-7B-0825-FP8fp8fp88.024,467allenai2025
Qwen_Qwen2.5-VL-7B-Instruct-GGUFggufgguf2.018,831bartowski2025
Qwen2.5-VL-7B-Instruct-GGUFggufgguf4.013,309ggml-org2025
Qwen2.5-VL-7B-Instructoriginalbf1616.013,034unsloth2025
Qwen2.5-VL-7B-Instruct-unsloth-bnb-4bitbnb-4bit4bit12,326unsloth2025
Qwen2.5-VL-7B-Instruct-bnb-4bitbnb-4bit4bit8,382unsloth2025
Qwen2.5-VL-7B-Instruct-GGUFggufgguf4.04,938mradermacher2025
GUI-Actor-7B-Qwen2.5-VLoriginalbf1616.03,909microsoft2025
Qwen2.5-VL-7B-Instruct-GGUFggufgguf16.03,595Mungert2025
olmOCR-7B-0825originalbf1616.02,634allenai2025
Qwen2.5-VL-7B-Instruct-FP8fp8fp88.01,006nvidia2025
X-Reasoner-7Boriginalbf1616.0838microsoft2026
olmOCR-7B-0725originalbf1616.0542allenai2025
Cosmos-Reason1-7B-GGUFggufgguf16.0278Mungert2025
V2P-7Boriginalbf1616.0209inclusionAI2025
Skywork-VL-Reward-7B-GGUFggufgguf16.0148Mungert2025
Holo1-7B-GGUFggufgguf16.0104Mungert2025
typhoon-ocr-7b-GGUFggufgguf16.094Mungert2025
Spatial-SSRL-7Boriginalbf1616.086internlm2026
olmOCR-7B-0725-FP8fp8fp88.078allenai2025
olmOCR-2-7B-1025-GGUFggufgguf16.075Mungert2025
ZwZ-7Boriginalbf1616.064inclusionAI2026
ViLaSRoriginalbf1616.042inclusionAI2025
Spark-VL-7Boriginalbf1616.039internlm2025
Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.