robit

ornith-vision

A self-improving family of open-source models for agentic coding, correctly exported with vision, reasoning and tool capabilities

vision tools thinking 9b 35b

602 Pulls 2 Tags Updated 2 weeks ago

qwen3.5-9b-r7-research-vision

Fine-tuned Qwen3.5-9B with distilled reasoning and full vision support. 883 tensors (427 text + 441 vision + 15 MTP) — vision tower preserved byte-for-byte from base via llama-export-lora merge.

vision tools thinking

594 Pulls 1 Tag Updated 3 months ago

ornith

A self-improving family of open-source models for agentic coding, correctly exported with reasoning and tool capabilities

tools thinking 9b 35b

513 Pulls 2 Tags Updated 3 weeks ago

qwen3.5-9b-r5-vision

Fine-tuned Qwen3.5-9B with distilled reasoning and full vision support. 883 tensors — vision tower preserved byte-for-byte from base. R5 was the first vision-capable distilled model.

vision tools thinking

461 Pulls 1 Tag Updated 3 months ago

qwen3.5-9b-r7-research

Fine-tuned Qwen3.5-9B with distilled reasoning from research-backed datasets. Trained via LoRA SFT with an additive data strategy that preserves base model capabilities while improving instruction following and reasoning.

tools thinking

127 Pulls 1 Tag Updated 3 months ago

qwen3.5-9b-r5-research

Fine-tuned Qwen3.5-9B with distilled reasoning from research-backed datasets. R5 was the first round to use production-quality data sources (Bespoke-Stratos, Tulu-3, SlimOrca) and achieved 84.2% on diverse eval — surpassing the base model.