Skip to content
Local AI Wizard
Your machine
Theme

Local vision models

Image to text (VLM) · Understand images and video: describe, answer questions, read documents, detect and segment.

Quality · LMArena

Human-preference rating on image questions · higher is better

  • Color = publisher

Source: LMArena · release of Oct 19, 2026 · checked Oct 25, 2026 · dataset · CC BY 4.0 · 143 groups have no rating in this release; they are left out, not scored as zero

Top rating

GLM-5.3-Flash: 1,271 on LMArena (vision)

Publishers

37 publishers in this task

Most common license

Apache 2.0 (89)

Memory estimates

134 of 162 groups, from Hub parameter counts

Choose your machine and the charts and tables mark what fits in its memory.

162 model groups · catalog checking Sep 26, 2026

Local vision models
GLM-5.3-Flashzai-org · MIT1,271 ± 6321240MITtransformersSep 7, 2026
gemma-4-31B-itgoogle · Apache 2.01,262 ± 631.324Apache 2.0transformers · safetensorsJul 20, 2026
Qwen3.8-27BQwen · Apache 2.01,242 ± 827.821Apache 2.0transformersAug 14, 2026
gemma-4-26B-A4B-itgoogle · Apache 2.01,242 ± 625.819Apache 2.0transformers · safetensorsJul 20, 2026
Qwen3-VL-235B-A22B-InstructQwen · Apache 2.01,215 ± 7236177Apache 2.0transformers · safetensorsNov 26, 2025
gemma-3-27b-itgoogle · Publisher terms: Gemma1,160 ± 827.421Publisher terms: Gemmatransformers · safetensorsMar 21, 2025
GLM-4.5Vzai-org · MIT1,153 ± 1210881MITtransformers · safetensorsOct 25, 2025
Llama-4-Maverick-17B-128E-Instructmeta-llama · Publisher terms1,145 ± 9402302Publisher termstransformers · safetensors · pytorchMay 22, 2025
step3stepfun-ai · Apache 2.01,145 ± 12321241Apache 2.0transformers · safetensorsJan 29, 2026
Llama-4-Scout-17B-16E-Instructmeta-llama · Publisher terms1,128 ± 1010982Publisher termstransformers · safetensors · pytorchMay 22, 2025
Qwen2.5-VL-32B-InstructQwen · Apache 2.01,119 ± 1533.526Apache 2.0transformers · safetensorsApr 14, 2025
Molmo2-8Ballenai · Apache 2.01,112 ± 208.77Apache 2.0transformers · safetensorsJan 23, 2026
Molmo-72B-0924allenai · Apache 2.01,045 ± 1373.355Apache 2.0transformers · safetensorsOct 9, 2025
Llama-3.2-90B-Vision-Instructmeta-llama · Publisher terms: Llama 3.21,032 ± 988.667Publisher terms: Llama 3.2transformers · safetensors · pytorchMar 4, 2025
Qwen2-VL-7B-InstructQwen · Apache 2.01,031 ± 118.37Apache 2.0transformers · safetensorsFeb 6, 2025
Molmo-7B-D-0924allenai · Apache 2.0994 ± 1486Apache 2.0transformers · safetensorsDec 15, 2025
Llama-3.2-11B-Vision-Instructmeta-llama · Publisher terms: Llama 3.2993 ± 1110.79Publisher terms: Llama 3.2transformers · safetensors · pytorchDec 4, 2024
MiniCPM-V-2_6openbmb962 ± 168.17Not declaredtransformers · safetensorsJun 13, 2025
Phi-3.5-vision-instructmicrosoft · MIT919 ± 164.14MITtransformers · safetensorsDec 10, 2025
Molmo2-4Ballenai · Apache 2.0Not on the LMArena vision leaderboard of 2026-10-194.94Apache 2.0transformers · safetensorsJan 23, 2026
MolmoMotion-4B-H1-F32allenai · Apache 2.0Not on the LMArena vision leaderboard of 2026-10-194.94Apache 2.0transformers · safetensorsJun 18, 2026
MolmoMotion-4B-H3-F30allenai · Apache 2.0Not on the LMArena vision leaderboard of 2026-10-194.94Apache 2.0transformers · safetensorsJun 18, 2026
olmOCR-2-7B-1025allenai · Apache 2.0Not on the LMArena vision leaderboard of 2026-10-198.37Apache 2.0transformers · safetensorsOct 22, 2025
olmOCR-2-7B-1025-FP8allenai · Apache 2.0Not on the LMArena vision leaderboard of 2026-10-198.3No estimate: a precision build of a listed model, packed weights, or a missing or partial Hub countApache 2.0transformers · safetensorsFeb 19, 2026
24 of 162