frob/ ornith-1.5:9b-mlx

73 yesterday

Missing quants, MTP, MLX and correct sampling parameters.

vision tools thinking 9b 35b 397b 
ollama run frob/ornith-1.5:9b-mlx

Details

yesterday

3b28627359f7 · 19GB ·

json
{ "architectures": [ "Qwen3_5ForConditionalGeneration" ], "dtype": "bfloat16", "eos_token_id": 24804
json
{ "size": { "longest_edge": 16777216, "shortest_edge": 65536 }, "patch_size": 16, "temporal_patch_si
json
{ "image_processor": { "do_convert_rgb": true, "do_normalize": true, "do_rescale": true, "do_resize"
json
{ "version": "1.0", "truncation": null, "padding": null, "added_tokens": [ { "id": 248044, "content"
json
{ "add_prefix_space": false, "added_tokens_decoder": { "248044": { "content": "<|endoftext|>", "lstr
json
{ "size": { "longest_edge": 25165824, "shortest_edge": 4096 }, "patch_size": 16, "temporal_patch_siz
json
{ "!": 0, "\"": 1, "#": 2, "$": 3, "%": 4, "&": 5, "'": 6, "(": 7, ")": 8, "*": 9, "+": 10, ",": 11,
license
MIT License Copyright (c) 2026 Ornith Team Permission is hereby granted, free of charge, to any pers
license
Copyright 2026 Alibaba Cloud Licensed under the Apache License, Version 2.0 (the "License"); you may
params
{ "min_p": 0, "presence_penalty": 1.5, "repeat_penalty": 1, "temperature": 1, "t
760 tensors

Readme

Imported from https://huggingface.co/collections/ornith-ai/ornith-15

MTP models imported from https://huggingface.co/protoLabsAI/Ornith-1.5-9B-MTP-GGUF.

-coding- models have parameters recommended by Ornith team for precise coding tasks.

Chat/thinking

$ ollama run frob/ornith-1.5:9b-mlx
>>> hello
Thinking...
The user just said "hello." This is a simple greeting. I should respond warmly 
and helpfully, perhaps offering to assist with whatever they need.

I don't need to overthink this. Just be friendly and open-ended so they can 
tell me what's on their mind.
...done thinking.

Hello! 👋 It's great to meet you.

How can I help you today? Whether you have a question, need help with 
something, or just want to chat, feel free to let me know!

>>> /clear
Cleared session context
>>> /set nothink
Set 'nothink' mode.
>>> hello
Hello! 👋 How can I help you today?

Tools

$ ollama launch opencode \
    --model frob/ornith-1.5:9b-mlx \
    -- \
    run 'create a file called README with the contents "hello world"'

> build · frob/ornith-1.5:9b-mlx

← Write README
Wrote file successfully.

Created README with the contents "hello world".

$ cat README 
hello world

Vision

$ ollama run frob/ornith-1.5:9b-mlx describe ./image1.jpg 
Added image './image1.jpg'
Thinking...
The user wants me to describe this image. Let me look at it carefully.

I see a small white puppy sitting on what appears to be a stone or concrete 
surface, possibly steps. The puppy has:
- Fluffy white fur
- Floppy ears
- Dark eyes
- A red collar with what looks like a bell (a golden/brass colored bell) 
attached
- There seem to be some thin whiskers visible on its face
- It's sitting upright in a cute pose, facing slightly to the right
- The background is blurred/dark - appears to be a doorway or entrance area

Let me write a clear, descriptive response.
...done thinking.

This image shows an adorable, fluffy white puppy sitting on what appears to be 
stone steps or a concrete surface.

**Key features:**
- **Fur:** Thick, soft-looking white fur covering its small body
- **Ears:** Floppy, rounded ears that sit close to its head
- **Face:** Dark, shiny eyes and a tiny black nose, with delicate whiskers 
visible on either side of its muzzle. The puppy is looking off to the right 
side of the frame.
- **Collar:** A red collar around its neck with a small golden/brass bell 
hanging down in front

**Setting:**
The puppy is sitting upright in an endearing pose. The background is softly 
blurred and dark, suggesting it may be near a doorway or entrance to a 
building. The surface it sits on appears to be a worn, textured stone step with 
visible speckling and wear.

The overall mood of the image is cute, gentle, and warm — a classic 
heartwarming shot of a young puppy.