vectionlabs/Salience-1.5-Nano

VISIONConcurrent Unit Cost:1Model Size:9BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jul 22, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

Vection Labs' Salience 1.5 Nano is a 9.65 billion parameter dense multimodal language model, built on Qwen3.5, designed for code-centric tasks. It features native vision and reasoning capabilities with a substantial 262,144 token context window. Optimized to run efficiently on a single consumer GPU, this model excels at generating and applying code patches, planning task sequences, and interpreting visual inputs like screenshots and diagrams.

Loading preview...

Salience 1.5 Nano: The 9B Model for Code-Centric Tasks

Salience 1.5 Nano, developed by Vection Labs, is a 9.65 billion parameter dense multimodal model built on the Qwen3.5 architecture. Unlike many smaller models that offer suggestions, Nano is engineered to actively work within a codebase, focusing on task completion rather than just providing information. It integrates native vision and reasoning, allowing it to process visual inputs and think autonomously without explicit prompting.

Key Capabilities

  • Code Patching: Identifies root causes of issues and generates minimal, effective code changes to fix them.
  • Terminal Integration: Capable of planning multi-step sequences, validating results at each stage, and recovering from failures.
  • Native Vision: Processes diverse visual inputs, including screenshots, diagrams, and stack traces, as first-class data.
  • Autonomous Reasoning: Features built-in reasoning capabilities, eliminating the need for "think step by step" prompts.
  • Efficient Deployment: Designed to run on a single consumer GPU, making it accessible for local development.

Good For

  • Developers needing an AI assistant that can directly interact with and modify codebases.
  • Automating debugging and code correction workflows.
  • Tasks requiring interpretation of visual development assets alongside text.
  • Environments where local deployment on consumer hardware is a priority.