The best vision and multimodal AI models in 2026

The best vision and multimodal AI models right now are SAM 3 (Segment Anything Model 3), Nemotron 3.5 Super VL and MiniCPM-V 4.7 (35B-A3B and 1B), out of 7 we have reviewed. Each is ranked by our rating, with its price, our verdict and who it suits.

By Andrew Miller · Updated Nov 2, 2026 · How we review

How we ranked: by our rating out of 5, from reviews built on each product's own pricing, documentation and policies plus independent benchmarks where they exist, each checked on the date shown in the review. We take no affiliate commission.

1. 1
   S

   SAM 3 (Segment Anything Model 3)

   SAM 3 turns "outline every forklift in this video" into one prompt, at more than twice the accuracy of the open-vocabulary detectors before it, though still well short of a human labeller.

   Best for: Computer-vision teams that need masks or boxes for named objects across images and video without training a detector for each class.

   Price: Free under the SAM License after a gated sign-up on Hugging Face; commercial use allowed with military, ITAR and reverse-engineering restrictions. Hosted use through Roboflow is billed on Roboflow's plans.

   Read our SAM 3 (Segment Anything Model 3) review
   4.4/5
2. 2
   

   Nemotron 3.5 Super VL

   The open model to beat at 120B on independent tests, provided the job is documents and research rather than code.

   Best for: Teams running document and research agents who want the strongest open mid-size model, especially on NVIDIA hardware.

   Price: Free weights under OpenMDW-1.1. Hosted: OpenRouter from $0.15 input / $0.50 output per 1M tokens, plus a free variant; NVIDIA NIM and other providers priced separately.

   Read our Nemotron 3.5 Super VL review
   4.1/5
3. 3
   

   MiniCPM-V 4.7 (35B-A3B and 1B)

   On OpenBMB's own numbers this is the open model to beat for reading documents and charts at the 35B size, and the 1B is the rare phone-sized vision model worth shipping.

   Best for: Teams extracting text and tables from scanned documents and charts on their own hardware, and mobile developers who need a vision model under 2 GB.

   Price: Free Apache 2.0 weights for both sizes. No first-party hosted price published; you pay for your own hardware.

   Read our MiniCPM-V 4.7 (35B-A3B and 1B) review
   4.0/5
4. 4
   M

   Mage-VL-Cowork

   Mage-VL-Cowork gets a 4B model within a few points of the best open desktop agents by letting it reach for a terminal, but Microsoft labels it research-only and Linux is all it has been tested on.

   Best for: Researchers and agent-platform teams studying small computer-use agents on Linux, who can run them in disposable sandboxes.

   Price: Free under Apache 2.0, though Microsoft's repository says the Mage models are for research and not intended for product or service deployment. No hosted API.

   Read our Mage-VL-Cowork review
   3.9/5
5. 5
   Z

   ZDTaichu5.0-9B

   A specialist dressed as a generalist: on its makers' own numbers it reads rooms and robot scenes far better than Qwen3.5-9B, and reads text slightly worse.

   Best for: Robotics and embodied-AI researchers who need a small open model that reasons about viewpoints, layouts and affordances, and can live with a forked serving stack.

   Price: Free weights under the NVIDIA Open Model License, with the Qwen3.5 parts under Apache 2.0. No hosted API or per-token price is published; you pay for your own GPU.

   Read our ZDTaichu5.0-9B review
   3.8/5
6. 6
   S

   SenseNova-Loop-9B-Preview

   SenseNova-Loop-9B-Preview posts agent scores well above its size, but it is a preview with no technical report, a 38 GB download and serving support still in unmerged branches.

   Best for: Research teams studying looped and weight-shared architectures, or evaluating small open models for office-agent work, who can tolerate preview-grade tooling.

   Price: Free under Apache 2.0. No hosted API listed for this model.

   Read our SenseNova-Loop-9B-Preview review
   3.5/5
7. 7
   

   Phi-4-multimodal

   Still the only Phi model that hears speech, but at 20 months old it has been overtaken by newer small multimodal models.

   Best for: Developers who want a small MIT-licensed model for on-device speech and image input.

   Price: Free weights under MIT. Azure AI Foundry: $0.08 input / $0.32 output per 1M text or image tokens; audio input $4 per 1M.

   Read our Phi-4-multimodal review
   3.3/5