Ornith 1.0 (9B)
DeepReinforce's Ornith 1.0 at 9B parameters (dense, built on Qwen 3.5 + Gemma 4), an agentic-coding model trained with a self-improving reinforcement-learning framework. Native <think> reasoning, OpenAI-style tool calling, 256K context. The small-tier pick for tool-heavy work on a 12GB+ inference budget. Q4_K_M quantization. MIT licensed.
At a glance
- Parameters
- 9B
- Quantization
- Q4_K_M
- Context window
- 262K tokens
- Approx. size
- 5.6 GB
- Engines
- llama.cpp, MLX
- License
- MIT (open)
- Version
- v1.0.0 , released 2026-07-01
- Family
- qwen
Tuning
Defaults a gezel client applies out of the box, from real evaluation runs against this exact quantization.
- sampling.temperature
0.6- sampling.topP
0.95- sampling.topK
20- sampling.minP
0- sampling.maxTokens
16384- reasoning.thinkingBudget
2048- reasoning.enableThinking
true- profiles
- thinking-general, thinking-coding, thinking-precise, instruct, creative
Model behaviors
Client-side behavior modules the catalog enables for this model (reasoning-tag handling, fabrication detection, prompt shaping).
- reasoning.strip-think-tags
- prompt.private-reasoning-guidance
- prompt.tool-cookbook-condensed
- fabrication.detect-past-tense-no-tools
- mcp.relax-required-fields
- mcp.default-missing-fields
- turn.ollama-num-predict-bumped
- turn.preamble-folding
- provider.merge-system-messages
Sources
- llama.cpp (GGUF)
- deepreinforce-ai/Ornith-1.0-9B-GGUF ·
ornith-1.0-9b-Q4_K_M.gguf· Q4_K_M - MLX
- mlx-community/Ornith-1.0-9B-4bit
- Upstream
- https://huggingface.co/deepreinforce-ai/Ornith-1.0-9B
- Manifest
- View on GitHub
Tags
- deepreinforce
- coding
- agentic
- tools
- reasoning
- long-context