Ornith-1.5 extends the self-scaffolding framework introduced in Ornith-1.0 into a more complete self-improvement loop:
The model proposes new tasks, generates task-specific scaffolds, and produces solution rollouts for reinforcement learning, continuously creating new learning experiences from which it can improve.

Ornith-1.5 9B is the direct successor of Ornith-1.0 9B, which achieves state-of-the-art performance among open-source models of comparable size across a broad range of agentic coding benchmarks.
It brings improved instruction following & improved thinking/reasoning, among other benefits.

[!TIP] Highly Recommended: Always keep reasoning/thinking enabled.
Ornith thoroughly plans and reasons through code edits before execution, ensuring an efficient and clean output.
Unlike baseline Qwen models, which frequently execute blindly and backtrack after generating broken code.

To learn more about Ornith 1.5, read their blog post.
To learn more about how to use Ornith 1.5 9B, view the base model.
A larger variant is also available: Ornith-1.5-35B-A3B

** Doesn't handle parallel tool calls well ** EDIT: It does. You have to enable reasoning.

--- Testing tool-calling model: ornith:9b ---

Testing non-stream mode...
L1 Basic Tool Call: What's the current weather in Paris right now? Use the tools available to you.
Check result: PASS (called get_current_weather with location 'Paris')
L2 Argument Extraction: A friend of mine lives in the city where the Eiffel Tower stands. Use your tools to tell me the current weather there.
Check result: PASS (called get_current_weather with location 'Paris')
L3 Tool Restraint: Do not call any tools. Answer directly from your own knowledge: what is the capital of France?
Check result: PASS (answered directly without any tool call)
L4 Parallel Tool Calls: Using your tools, compare the current weather in Paris and Tokyo.
Check result: PASS (2 parallel calls covering both cities)
L5 Multi-Turn Tool Loop: Use the price lookup tool to check the unit price of a 'widget', then tell me what 3 widgets would cost after a 10% discount. Do the math yourself.
Lookup check result: PASS (requested price for 'widget')
Final answer check result: PASS (final answer contains the computed total ($54))

This repository contains a quantized Q4NX port of the model, compiled for the FastFlowLM (FLM) runtime. It is not a GGUF file.

Item	Value
Source model	`ornith-ai/Ornith-1.5-9B`
Source GGUF	`Ornith-1.5-9B-MXFP4_Q8_0-Imatrix.gguf`
Weights	`model.q4nx` (7.11 GB)
Modality	language / vision
FLM version	`1.0.1`
Converted	2026-08-24

Install and run

This repository works with flm-add, a small installer that copies the model into the FastFlowLM user directory and registers the tag. It never modifies the system FastFlowLM install.

pip install flm-add or uv tool install flm-add

bash

uv tool install flm-add
flm-add Atomic-Germ/Ornith-1.5-9B-NPU2 --family qwen3.5 --tag ornith1.5:9b
FLM_CONFIG_PATH="$HOME/.config/flm/model_list.json" FLM_XCLBIN_PATH="$HOME/.config/flm" flm run ornith1.5:9b

Files

File	Description
`model.q4nx`	Quantized weights (Q8_0 / Q4_1 / BF16)
`config.json`	FLM runtime configuration
`tokenizer.json`	Tokenizer vocabulary
`tokenizer_config.json`	Tokenizer configuration
`chat_template.jinja`	Chat template
`vision_weight.q4nx`	Vision model

Source model card

See the original model card: ornith-ai/Ornith-1.5-9B