A July 2026 discussion on choosing local VLMs
This is a July 2026 discussion about which local VLMs are worth using now. The comparison is limited to models. Useful answers need more than s; they should include the hardware used, the , the real use case, how often the model is used, and whether the work is personal or professional.
VLM is treated as difficult because can be unreliable, tools are still immature, and results can vary between runs. Details about tools, , and prompts are needed so other people can judge whether a recommendation applies to their own setup.
Key points
- The focus is local VLMs available in July 2026.
- Only models are allowed in the comparison.
- Recommendations should include hardware, , use case, and usage level.
- alone may not be trustworthy for VLMs.
- For AI agents, real setup details matter more than a .