Skip to content

EthenEthenEthen

LoginPlatformTry Ethen

Open Source Model Profile · vectionlabs

Salience-1-9B

Salience-1-9B is an 8.77B-parameter vision-language model from vectionlabs. According to the model card, it targets code, agentic tool use, math reasoning, and image and video understanding with up to 1M tokens of context.

Publisher
vectionlabs
Task
image-text-to-text
Model type
qwen3_vl
License
apache-2.0
Library
transformers
Publication status
Approved for indexing

Model overview

Salience-1-9B is published by vectionlabs as an image-text-to-text model. The captured configuration identifies Qwen3VLForConditionalGeneration with model type qwen3_vl, and Safetensors metadata reports 8,767,123,696 parameters. According to the model card, it is a dense 9B model built on Qwen3-VL and presented as the successor of Maestro1-9B.

Recorded capabilities

Qwen3-VL architecture

According to the model card, the model couples a 36-layer Qwen3-8B language model with a native vision encoder using interleaved multimodal RoPE.

1M-token context claim

According to the model card, context extends from 256K up to 1M tokens for whole repos, long papers, or long videos.

Code and agentic emphasis

According to the model card, the release prioritizes runnable code, repo-scale edits, tool calls, math reasoning, and visual understanding.

Single-GPU deployment notes

According to the model card, bf16/fp16 loads with device_map auto on one modern accelerator, while 4-bit quantization fits a single consumer GPU.

Use cases in the source record

  • Code generation, explanation, debugging, review, and repo-scale editing, as listed in the card's intended-use section.
  • Agentic and tool-use workflows requiring runnable code and well-formed tool calls, as described in the card's capability pillars.
  • Long-context multimodal work over repos, papers, images, and video within the card-reported 1M-token context.

Limitations and unknowns

  • Context-window, benchmark, and capability statements are publisher claims and were not independently verified.
  • No evaluation results were extracted as structured data from this record.
  • Provider state is historical snapshot data, not independently refreshed current availability.

Source and provenance

Source: vectionlabs/Salience-1-9B

Captured: Unknown. Processed: 2026-09-07T19:36:10.810276+00:00.

Salience 1 — 9B A 9B multimodal reasoning model, sharpened for code and agentic work — that can see. Vection Labs Weights · Benchmarks · Quickstart · Fast inference · Limitations Abstract Salience 1 (9B) is a dense, 9-billion-parameter vision-language model built for hard, practical work : writing and debugging real code, driving tools and agents, multi-step mathematical reasoning, and visual understanding over images and video — inside a single model with a context window of up to 1M tokens . It is the successor of Maestro1-9B , engineered around a single goal: push the axis users ask for most — code and agentic/tool use — without…

F001F002F003F004F005F006F007F008F009F010F011F013F014F016F017F018F019F020F021F022