Instructions to use netprojects/DaisyRose-27B-MLX-2.5bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use netprojects/DaisyRose-27B-MLX-2.5bit with MLX:
# Make sure mlx-lm is installed # pip install --upgrade mlx-lm # Generate text with mlx-lm from mlx_lm import load, generate model, tokenizer = load("netprojects/DaisyRose-27B-MLX-2.5bit") prompt = "Write a story about Einstein" messages = [{"role": "user", "content": prompt}] prompt = tokenizer.apply_chat_template( messages, add_generation_prompt=True ) text = generate(model, tokenizer, prompt=prompt, verbose=True) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Pi
How to use netprojects/DaisyRose-27B-MLX-2.5bit with Pi:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "netprojects/DaisyRose-27B-MLX-2.5bit"
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "mlx-lm": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "netprojects/DaisyRose-27B-MLX-2.5bit" } ] } } }Run Pi
# Start Pi in your project directory: pi
- MLX LM
How to use netprojects/DaisyRose-27B-MLX-2.5bit with MLX LM:
Generate or start a chat session
# Install MLX LM uv tool install mlx-lm # Interactive chat REPL mlx_lm.chat --model "netprojects/DaisyRose-27B-MLX-2.5bit"
Run an OpenAI-compatible server
# Install MLX LM uv tool install mlx-lm # Start the server mlx_lm.server --model "netprojects/DaisyRose-27B-MLX-2.5bit" # Calling the OpenAI-compatible server with curl curl -X POST "http://localhost:8080/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "netprojects/DaisyRose-27B-MLX-2.5bit", "messages": [ {"role": "user", "content": "Hello"} ] }' - Hermes Agent
How to use netprojects/DaisyRose-27B-MLX-2.5bit with Hermes Agent:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "netprojects/DaisyRose-27B-MLX-2.5bit"
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:8080/v1 hermes config set model.default netprojects/DaisyRose-27B-MLX-2.5bit
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use netprojects/DaisyRose-27B-MLX-2.5bit with OpenClaw:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "netprojects/DaisyRose-27B-MLX-2.5bit"
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:8080/v1 \ --custom-model-id "netprojects/DaisyRose-27B-MLX-2.5bit" \ --custom-provider-id mlx-lm \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
DaisyRose-27B for Apple Silicon (MLX, 2.5-bit)
RadixArk/DaisyRose-27B (revision 293a7a2) converted to MLX with mlx-vlm,
DaisyRose's own multi-token-prediction (MTP) layer included. This is a text-only build: the vision encoder is removed. The MLP gate and up projections are 2-bit, the down projections 3-bit, every attention and linear-attention (Gated DeltaNet) projection 4-bit, and the token embedding and output layer 4-bit (affine, group 64), about 3.65 bits per weight overall. Size: 11.8 GB.
DaisyRose-27B is RadixArk's reinforcement-learned version of Qwen3.8-27B: trained with RL only (no supervised fine-tuning or distillation) on about 1,200 long-horizon agent tasks. It keeps Qwen3.8-27B's architecture, tokenizer, chat template, image understanding and 256K context, and thinks before it answers by default. RadixArk's published results against the base model:
| Benchmark | DaisyRose-27B | Qwen3.8-27B |
|---|---|---|
| Terminal Bench 4.1 | 12.0 | 5.6 |
| Terminal Bench 2.1 | 75.5 | 73.0 |
| DeepSWE v1.1 | 46.2 | 42.2 |
| SWE Bench Pro v2 | 94.7 | 91.3 |
| HLE (text only) | 34.3 | 33.9 |
| IFBench | 82.4 | 79.5 |
| GPQA Diamond | 90.7 | 89.2 |
Publisher's figures for the original BF16/FP8 weights; see the original model card for the harnesses, repetitions and other models.
| Build | Size | Bits per weight | Same next token as bf16 | KL vs bf16 | GSM8K, 50 questions | Peak memory, 8K-token prompt | Recommended Mac | Tool call / coding / thinking / vision |
|---|---|---|---|---|---|---|---|---|
| 6-bit (recommended) | 23.1 GB | 6.66 | 98.7% | 0.0014 | 44/50 | 23.4 GB | 36 GB+ | pass / 3/3 / 2/2 / pass |
| 8-bit | 29.9 GB | 8.63 | 99.5% | 0.0004 | 45/50 | 29.7 GB | 48 GB+ | pass / 3/3 / 2/2 / pass |
| 4-bit | 16.3 GB | 4.70 | 95.8% | 0.0207 | 45/50 | 17.1 GB | 32 GB+ | pass / 3/3 / 2/2 / pass |
| 3-bit | 13.8 GB | 3.91 | 92.1% | 0.0814 | 43/50 | 14.6 GB | 24 GB+ | pass / 3/3 / 2/2 / pass |
| 2.5-bit | 11.8 GB | 3.65 | 87.6% | 0.2138 | 42/50 | 13.8 GB | 24 GB+ | pass / 2/3 / 2/2 / text-only |
| bf16 | 55.6 GB | 16 | reference | โ | โ | 53.4 GB | 96 GB+ | โ / โ / โ / pass |
Measured on our builds as described under "What was checked" (GSM8K: thinking off, greedy; peak memory: mlx-lm, text). "Recommended Mac" is the smallest Mac whose default GPU memory limit (about two thirds of RAM up to 36 GB, three quarters above) holds that peak; longer contexts need more.
On a smaller Mac you can raise the GPU memory limit until the next restart, for example sudo sysctl iogpu.wired_limit_mb=27000 on a 32 GB Mac
(leave several GB for macOS and your apps). With a raised limit and short contexts the next size down usually works: 6-bit on 32 GB, 4-bit on 24 GB. The 6-bit build is the recommended balance of size and fidelity; 8-bit is closest to bf16; 4-bit, 3-bit and the text-only 2.5-bit are for Macs with less memory (3-bit and 2.5-bit fit 24 GB Macs).
Use with mlx-lm
The model thinks by default (the thinking comes back before </think>). RadixArk's recommended sampling for thinking mode is
temperature 1.0, top_p 0.95, top_k 20, min_p 0; leave room for long thinking on hard problems.
pip install -U "mlx-lm>=0.32.2"
mlx_lm.server --model netprojects/DaisyRose-27B-MLX-2.5bit --port 8000 --temp 1.0 --top-p 0.95 --top-k 20
from mlx_lm import load, generate
from mlx_lm.sample_utils import make_sampler
model, tokenizer = load("netprojects/DaisyRose-27B-MLX-2.5bit")
messages = [{"role": "user", "content": "How many positive divisors does 360 have?"}]
prompt = tokenizer.apply_chat_template(messages, add_generation_prompt=True, tokenize=False)
print(generate(model, tokenizer, prompt=prompt, max_tokens=8192, sampler=make_sampler(temp=1.0, top_p=0.95, top_k=20)))
The chat template takes reasoning_effort (xhigh default, medium, low) and enable_thinking=False to answer
directly. Tool calls use Qwen's XML format (<tool_call><function=...><parameter=...>), which mlx-lm's server returns as
OpenAI tool_calls. The tokenizer and chat template are RadixArk's files, unchanged.
Faster decoding with MTP
This build keeps DaisyRose-27B's own MTP layer in model-mtp.safetensors, quantized to 4-bit. mlx-vlm uses it as a built-in draft
model when you pass the same repository as --draft-model:
mlx_vlm.generate --model netprojects/DaisyRose-27B-MLX-2.5bit --draft-model netprojects/DaisyRose-27B-MLX-2.5bit --draft-kind mtp \
--temperature 0 --max-tokens 512 --prompt "Write a Python function that checks whether a number is prime."
Every drafted token is checked by the full model. With greedy decoding the text matches plain decoding until a near-tie, where the batched check can round the other way, so long answers can diverge slightly.
What was checked
On a MacBook Pro M5 Max (128 GB), 19 November 2026, with mlx 0.32.4, mlx-lm 0.32.2 and mlx-vlm 0.7.11.
- Source. Every file of RadixArk/DaisyRose-27B at revision
293a7a2matches the Hub's SHA-256 hashes; its config and tensor list are identical to Qwen3.8-27B's. - Quantization. Each quantized build is checked to be exactly the affine (group 64) quantization of the bf16 build, with the bit widths described above.
- Agreement with bf16 (12 short prompts, 9 with thinking off and 3 with thinking on, 953 teacher-forced positions): share of positions where each build picks the same next token as bf16: 8-bit 99.5% (KL 0.0004); 6-bit 98.7% (KL 0.0014); 4-bit 95.8% (KL 0.0207); 3-bit 92.1% (KL 0.0814); 2.5-bit 87.6% (KL 0.2138).
- Tool calling and coding (thinking off): one agent-style prompt with
bashandreadtools and three short Python tasks run against asserts: 8-bit tool call passed, coding 3/3; 6-bit tool call passed, coding 3/3; 4-bit tool call passed, coding 3/3; 3-bit tool call passed, coding 3/3; 2.5-bit tool call passed, coding 2/3. - Thinking (two short maths prompts, reasoning effort low): the reply closes
</think>once and the answer is right: 8-bit 2/2; 6-bit 2/2; 4-bit 2/2; 3-bit 2/2; 2.5-bit 2/2. - GSM8K, 50 fixed test questions (thinking off, greedy, 512-token cap): 8-bit 45/50, 6-bit 44/50, 4-bit 45/50, 3-bit 43/50, 2.5-bit 42/50.
- Vision (mlx-vlm): asked to name the shapes and colours in a test image, every build with vision (all but the text-only 2.5-bit) answered correctly and answered a text-only question in the same session.
- MTP layer. The bundled MTP tensors equal DaisyRose-27B's
mtp.*weights (norms in mlx-vlm's stored form), quantized exactly. Every quantized build was loaded in mlx-lm and mlx-vlm with the extra file present and decoded with its bundled MTP head (the 3-bit build keeps an 8-bit head and the 2.5-bit build a 4-bit head, declared inconfig.jsonasmtplx_mtp_quantization).
Credits
DaisyRose-27B is by RadixArk; the base model is Qwen3.8-27B by the Qwen Team, Alibaba Cloud. If you use the model, please cite their work (see the DaisyRose-27B card and the Qwen3.8-27B card).
License
Apache-2.0, as DaisyRose-27B and Qwen3.8-27B. LICENSE is the Apache-2.0 text from the DaisyRose-27B repository and
NOTICE describes the conversion and credits RadixArk and the Qwen Team (Alibaba Cloud).
Published by NetProjects (Internet Projects Ltd). Not affiliated with RadixArk or Alibaba.
More MLX builds from NetProjects: huggingface.co/netprojects ยท netprojects.com
- Downloads last month
- -
2-bit