Jan-v2-VL-f32-AIO-GGUF

Jan-v2-VL-f32-AIO-GGUF

Jan-v2-VL is an 8-billion-parameter vision-language model designed for long-horizon, multi-step task execution in real software environments such as browsers and desktop applications. It tightly integrates language reasoning with visual perception to follow complex instructions, maintain intermediate states, and recover from minor execution errors, enabling stable and continuous task completion. The model is evaluated with a benchmark focused on long-horizon execution, confirming its strength in steady, low-drift stepwise operation. Jan-v2-VL comes in three variants—low (efficiency and lower latency), medium (balanced latency and quality), and high (deeper reasoning with higher think time)—to best suit different performance needs. It supports agentic automation and UI control via screenshot grounding and tool calls and is optimized for local deployment with popular frameworks like vLLM and llama.cpp, offering a powerful tool for automation and interaction tasks in complex visual and textual environments.

Models

Jan-v2-VL-high

Jan-v2-VL-med

Jan-v2-VL-low

Quants Usage

(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)

Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):