HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks

Phi-3.5-vision-instruct

microsoft/Phi-3.5-vision-instruct

kind
original
quant
bf16 (16-bit)
family
phi
modality
image-text-to-text
root model
microsoft/Phi-3.5-vision-instruct
provider
microsoft
license
mit
language
multilingual
context
128K
library
transformers
vRAM (GB)
—
size on disk
—
downloads
1,232,249
likes
737
created
—
last modified
2025-12-10T20:24:07+00:00 (2025)
huggingface
https://huggingface.co/microsoft/Phi-3.5-vision-instruct
tags
transformerssafetensorsphi3_vtext-generationnlpcodevisionimage-text-to-textconversationalcustom_codemultilingualarxiv:2404.14219license:mitregion:us

Benchmark scores (1)

benchmarkcategoryscorecontextconfevidence
MATHmath43.9Visual reasoningmedium| Visual math reasoning | MathVista (testmini) | 43.9 | 53.7 | 51.1 | 55.3 | 38.8 | 54.0 | 57.4 | 54.4 |

Discussions (0)

none scraped.

Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.