lmstudio-community / qwen-2.5 family / image-text-to-text
qwen2.5-vl-72b
qwen/qwen2.5-vl-72b
Run locally Smallest indexed GGUF / needs Ollama
ollama run hf.co/qwen/qwen2.5-vl-72b/Qwen_Qwen2.5-VL-72B-Instruct-IQ1_M.gguf Benchmarks
Where it scores.
General Provisional
60.6
Coding Provisional —
Math Provisional —
Reasoning Provisional
51.1
Normalized 0–100 across indexed models.
Runnable variants
Pick a size.
| Precision | Format | Size | Params | File |
|---|---|---|---|---|
| IQ1_M / smallest | GGUF | 23.7 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-IQ1_M.gguf |
| IQ2_XXS | GGUF | 25.5 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-IQ2_XXS.gguf |
| IQ2_XS | GGUF | 27.1 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-IQ2_XS.gguf |
| IQ2_S | GGUF | 27.9 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-IQ2_S.gguf |
| IQ2_M | GGUF | 29.3 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-IQ2_M.gguf |
| Q2_K | GGUF | 29.8 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q2_K.gguf |
| Q2_K_L | GGUF | 31.0 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q2_K_L.gguf |
| IQ3_XXS | GGUF | 31.8 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-IQ3_XXS.gguf |
| IQ3_XS | GGUF | 32.8 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-IQ3_XS.gguf |
| Q3_K_S | GGUF | 34.5 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q3_K_S.gguf |
| IQ3_M | GGUF | 35.5 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-IQ3_M.gguf |
| Q3_K_M | GGUF | 37.7 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q3_K_M.gguf |
| Q3_K_L | GGUF | 39.5 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q3_K_L.gguf |
| IQ4_XS | GGUF | 39.7 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-IQ4_XS.gguf |
| Q3_K_XL | GGUF | 40.6 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q3_K_XL.gguf |
| IQ4_NL | GGUF | 41.3 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-IQ4_NL.gguf |
| Q4_0 | GGUF | 41.4 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q4_0.gguf |
| Q4_K_S | GGUF | 43.9 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q4_K_S.gguf |
| Q4_1 | GGUF | 45.7 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q4_1.gguf |
| Q4_K_M | GGUF | 47.4 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q4_K_M.gguf |
| Q4_K_L | GGUF | 48.3 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q4_K_L.gguf |
| Q5_K_S | GGUF | 51.4 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q5_K_S/Qwen_Qwen2.5-VL-72B-Instruct-Q5_K_S-00001-of-00002.gguf |
| Q5_K_M | GGUF | 54.4 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q5_K_M/Qwen_Qwen2.5-VL-72B-Instruct-Q5_K_M-00001-of-00002.gguf |
| Q6_K | GGUF | 64.3 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q6_K/Qwen_Qwen2.5-VL-72B-Instruct-Q6_K-00001-of-00002.gguf |
| Q8_0 | GGUF | 77.3 GB | — | Qwen_Qwen2.5-VL-72B-Instruct-Q8_0/Qwen_Qwen2.5-VL-72B-Instruct-Q8_0-00001-of-00002.gguf |
Sizes are file sizes, not peak runtime memory.
Raw benchmarks
The inputs.
| Source | Task | Raw | Normalized |
|---|---|---|---|
| ZeroEval | MMMU | 70.2 | 70.2 |
| ZeroEval | MMMUPro | 51.1 | 51.1 |
From the Hugging Face model card
About this model.
From the Hugging Face model card