Meta 开源 Muse Glimmer:30B 本地多模态智能体模型

原标题:Meta is back with Muse Glimmer: local, agentic, multimodal, and open source
86AI 研判分

Meta 发布并开源了专为本地智能体场景设计的多模态模型 Muse Glimmer,参数量蒸馏至 30B,采用 Apache 2.0 许可证。该模型旨在平衡隐私保护与成本效益,支持代码生成、文档分析及个人助理等应用,并已获 Hugging Face 生态(Transformers, vLLM 等)的 Day-0 全面适配。

报道正文

本文目录25 个章节

返回文章列表

Meta 携 Muse Glimmer 回归:本地化、智能体、多模态且开源!

发布于 2026 年 8 月 10 日 在 GitHub 上更新 点赞 116

Pedro Cuenca pcuenq merve merve ben burtenshaw burtenshaw Aritra Roy Gosthipaty ariG23498

来自开源 LLM 先驱的好消息!今天发布的 Muse Glimmer 是 Meta 推出的全新多模态模型,专为本地智能体(agentic)应用场景设计。该模型从 Muse 蒸馏至 30B 参数规模,并以 Apache 2.0 许可证发布,非常适合用于注重隐私的本地部署、降低成本或纯粹的技术探索。它旨在服务于隐私敏感型应用,例如编程、文档分析、个人助理以及类似 Claw 或 Hermes 的配置方案。

为庆祝这一发布,我们与 Meta 合作,在 transformers、llama.cpp、vLLM、Inference Endpoints 及其他库中提供了 Day-0 支持。我们构建了一些有趣的功能,并在本文中分享我们的发现。

查看下方的演示以获取灵感。

你可以在 Hugging Face Hub 上找到 Muse Glimmer。

基准测试

基准测试结果

分数按原始发布数据报告。加粗表示在对比模型中的最佳结果;↓ 表示数值越低越好。

类别BenchmarkMuse Glimmer-30B High ReasoningGemma4-31B Thinking ModeQwen3.6-27B Thinking Mode
通用智能体MCP Atlas75.554.262.5
通用智能体DeepSearch QA74.661.771.1
通用智能体τ³-Banking23.515.116.7
通用智能体WildClawBench47.637.643.2
通用智能体GDPval-AA9538111141
通用智能体GAIA243.336.440.0
通用智能体SkillsBench (With Skills)44.332.446.6
通用智能体OSWorld-Verified65.958.575.6
智能体编程SWE-Bench Pro51.236.950.2
智能体编程SWE-Bench Verified76.066.677.2
智能体编程TerminalBench 2.151.743.460.7
智能体编程SciCode43.643.439.8
多模态Charxiv Reasoning78.877.778.4
多模态ScreenSpot Pro75.475.976.1
多模态OmniDocBench v1.575.872.577.8
多模态MMMU Pro747375
安全性CI MemoriesViolation (↓): 26.4 Coverage: 64.8Violation (↓): 12.1 Coverage: 53.0Violation (↓): 53.4 Coverage: 66.9
安全性Siren AgentDojoAttack Success Rate (↓): 28.4 Utility: 94.2Attack Success Rate (↓): 25.6 Utility: 90.8Attack Success Rate (↓): 40.3 Utility: 92.7
通用能力与推理IFBench77.076.070.8
通用能力与推理AIME 202694.789.294.1
通用能力与推理GPQA Diamond83.585.784.2
通用能力与推理Humanity’s Last Exam (Text + No Tools)22.023.623.1
通用能力与推理AA-LCR80.068.373.3
通用能力与推理Beam 128K65.158.263.0

架构

Muse Glimmer 是一个包含 300 亿参数的稠密模型,由以下部分组成:

  • 用于视觉的 20 亿参数 ViT 风格编码器(感知编码器)
  • 280 亿参数的文本解码器

除了主 VLM 之外,还有一个基于 DFlash 的推测解码草稿模型。该模型的使用是可选的,可以在牺牲一定内存开销的前提下显著提升生成速度。我们发现该草稿模型在结构化内容生成任务中表现尤为出色。

文本解码器

  • Hybrid attention: Alternating between three sliding window layers (of 2,048 tokens) using rotary position embedding, followed by a fourth layer that uses full attention and NoPE (no positional embedding). The pattern is therefore (SWA, SWA, SWA, Full), repeated 13 times to a total of 52 layers. This allows the model to retain relative order and distance information with RoPE and preserve information globally with NoPE.
  • Gated Grouped-Query Attention: Each key-value head is shared by 16 query heads, which reduces KV-cache memory by 16x and makes generation faster and cheaper.
  • Q-K normalization with extra query scaling: Before computing attention, Muse Glimmer applies RMS normalization to every query and key head to keep attention logits stable. After this, queries are multiplied by a scale factor to set the target logit scale after normalization. The extra query scaling behaves like an inverse temperature at the softmax level.

Perception Encoder

Muse Glimmer uses one image encoder to handle both images and videos. Unlike the relatively small vision encoders used in other VLMs, this is a sizable 2B ViT-like model designed after the Perception Encoder architecture. Perception Encoder was previously introduced by Meta as a backbone for various downstream spatial and multimodal tasks. The encoder patchifies images to a shape of 2 frames x 3 channels x 14 x 14, and passes them through a linear layer for projection. An interpolated absolute position embedding from a learned position table is then added to these embeddings. These are then sent to the vision tower which consist of 50 layers and GELU MLPs. Similar to the language model, the attention pattern consists of three window attention layers followed by one full attention layer. Inside the attention layers, 2D RoPE is applied to the queries and keys.

After transformer, pixel shuffle concatenates 2x2 groups of neighboring spatial tokens which reduces the number of image tokens 4x without discarding their channels. The merged features are then projected to the shared embedding space of the text decoder.

Videos go through the same encoder frame by frame, where each frame is converted into patches (of shape [batch, temporal groups, grid height, grid width, 2 frames, 3 channels, 14, 14]). The processor targets 2 frames per second and caps the clip at 96 frames sampled evenly across video. The processor creates timestamped video placeholders, interleaving text with frame e.g. “Time: 0.0s <|video|> x N” in which the final video embeddings are replaced before the final projection layer.

Transformers

Upgrade transformers to the latest version to be able to use Muse Glimmer.

BASH
pip install --upgrade transformers accelerate

Muse Glimmer comes with day-0 support in transformers, both for the main model and the speculative decoding drafter. You can use AutoModelForMultimodalLM and AutoProcessor classes to load the model and the processor.

PYTHON
from transformers import AutoProcessor, AutoModelForMultimodalLM

MODEL_ID = "meta-models/Muse-Glimmer-30B"

processor = AutoProcessor.from_pretrained(MODEL_ID)
model = AutoModelForMultimodalLM.from_pretrained(
    MODEL_ID,
    dtype="auto",
    device_map="auto"
)

The same snippet runs unchanged on NVIDIA (CUDA), AMD (ROCm) and Intel (XPU) GPUs, device_map="auto" places the model on whichever accelerator is available.

Text-only Inference

After loading the model, you can do text-only inference with it as follows.

PYTHON
from transformers import AutoProcessor, AutoModelForMultimodalLM

MODEL_ID = "meta-models/Muse-Glimmer-30B"

processor = AutoProcessor.from_pretrained(MODEL_ID)
model = AutoModelForMultimodalLM.from_pretrained(
    MODEL_ID,
    dtype="auto",
    device_map="auto"
)

messages = [
    {"role": "user", "content": "Write a short joke about saving RAM."},
]

inputs = processor.apply_chat_template(
    messages,
    tokenize=True,
    return_dict=True,
    return_tensors="pt",
    add_generation_prompt=True,
    reasoning_strength="low"
).to(model.device)
input_len = inputs["input_ids"].shape[-1]

outputs = model.generate(**inputs)
response = processor.decode(outputs[0][input_len:], skip_special_tokens=False)
print(response)

Prompting the model with images and text

We would need torchvision to be able to use images and text.

BASH
pip install torchvision

Muse Glimmer accepts images as input, as demonstrated here:

PYTHON
from transformers import AutoProcessor, AutoModelForMultimodalLM

MODEL_ID = "meta-models/Muse-Glimmer-30B"



> *[注:本文篇幅超长,以上为核心前篇精译,后续内容保留原文呈现]*


processor = AutoProcessor.from_pretrained(MODEL_ID)
model = AutoModelForMultimodalLM.from_pretrained(
    MODEL_ID,
    dtype="auto",
    device_map="auto"
)

messages = [
    {
        "role": "user", "content": [
            {"type": "image", "image": "https://huggingface.co/datasets/merve/vl-test-suite/resolve/main/SF.png"},
            {"type": "text", "text": "What is shown in this image?"}
        ]
    }
]

inputs = processor.apply_chat_template(
    messages,
    tokenize=True,
    return_dict=True,
    return_tensors="pt",
    add_generation_prompt=True,
    reasoning_strength="low"
).to(model.device)
input_len = inputs["input_ids"].shape[-1]

outputs = model.generate(**inputs)
response = processor.decode(outputs[0][input_len:], skip_special_tokens=False)
print(response)

Video Inference

To work with videos we recommend installing torchcodec into the environment.

BASH
pip install torchcodec

Muse Glimmer can answer complex questions about videos without audio. You can do video inference as follows, here’s an example from VideoMME2, which is the most popular video question answering benchmark.

PYTHON
from transformers import AutoProcessor, AutoModelForMultimodalLM

MODEL_ID = "meta-models/Muse-Glimmer-30B"

processor = AutoProcessor.from_pretrained(MODEL_ID)
model = AutoModelForMultimodalLM.from_pretrained(
    MODEL_ID,
    dtype="auto",
    device_map="auto"
)

messages = [
    {
        "role": "user",
        "content": [
            {"type": "video", "video": "https://huggingface.co/datasets/merve/vl-test-suite/resolve/main/IMG_8137.mp4"},
            {"type": "text", "text": "Describe what happens in this video."},
        ],
    },
]

inputs = processor.apply_chat_template(
    messages,
    tokenize=True,
    return_dict=True,
    return_tensors="pt",
    add_generation_prompt=True,
    reasoning_strength="low",
    processor_kwargs={"num_frames": 96},
).to(model.device)

input_len = inputs["input_ids"].shape[-1]
outputs = model.generate(**inputs)

response = processor.decode(
    outputs[0, input_len:],
    skip_special_tokens=False,
)
print(response)

Multimodal tool calling

Muse Glimmer can do multimodal tool calling, here’s how you can do it. In the example below, we ask the model to call the weather tool based on the city in the image.

PYTHON
import json
import re
from transformers import AutoProcessor, AutoModelForMultimodalLM

MODEL_ID = "meta-models/Muse-Glimmer-30B"

processor = AutoProcessor.from_pretrained(MODEL_ID)
model = AutoModelForMultimodalLM.from_pretrained(
    MODEL_ID,
    dtype="auto",
    device_map="auto"
)

tools = [
    {
        "type": "function",
        "function": {
            "name": "weather.get",
            "description": "Get the current weather for a city.",
            "parameters": {
                "type": "object",
                "properties": {
                    "city": {"type": "string"},
                },
                "required": ["city"],
            },
        },
    }
]

messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "image": "https://huggingface.co/datasets/merve/vl-test-suite/resolve/main/SF.png"},
            {"type": "text", "text": "I'm going to the city in this picture. What clothes should I wear?"},
        ],
    },
]

inputs = processor.apply_chat_template(
    messages,
    tools=tools,
    tokenize=True,
    return_dict=True,
    return_tensors="pt",
    add_generation_prompt=True,
    reasoning_strength="low"
).to(model.device)

input_len = inputs["input_ids"].shape[-1]
outputs = model.generate(**inputs)
response = processor.decode(outputs[0][input_len:], skip_special_tokens=False)
print(response)

Object Detection

You can use Muse Glimmer to do open ended object detection in images as follows.

PYTHON
from transformers import AutoProcessor, AutoModelForMultimodalLM

MODEL_ID = "meta-models/Muse-Glimmer-30B"

processor = AutoProcessor.from_pretrained(MODEL_ID)
model = AutoModelForMultimodalLM.from_pretrained(
    MODEL_ID,
    dtype="auto",
    device_map="auto"
)

messages = [{
    "role": "user",
    "content": [
        {"type": "image", "image": "https://huggingface.co/datasets/merve/vl-test-suite/resolve/main/SF.png"},
        {
            "type": "text",
            "text": (
                "Detect the bridge. Return only the detection in the model's "
                "native object-detection format, with no explanation."
            ),
        },
    ],
}]

inputs = processor.apply_chat_template(
    messages,
    tokenize=True,
    return_dict=True,
    return_tensors="pt",
    add_generation_prompt=True,
    reasoning_strength="low",
).to(model.device)

input_len = inputs["input_ids"].shape[-1]
outputs = model.generate(**inputs, max_new_tokens=128)
response = processor.decode(outputs[0][input_len:], skip_special_tokens=False)

detections = json.loads(response.removesuffix("<|eot|>"))
print(detections)

Here is an end to end script to perform object detection GitHub Gist

Llama.cpp

Muse Glimmer comes with day-0 llama.cpp support. Meta has distributed calibrated quants in this repo, and Unsloth is releasing optimized quants as well. DFlash speculative decoding is supported as well. You can use a pre-built llama binary to start a llama server or a CLI. To install llama.cpp, run:

BASH
curl -LsSf https://llama.app/install.sh | sh

Then you can start the server as follows.

BASH
llama serve -hf meta-models/Muse-Glimmer-30B-GGUF

Once the server has started, you can head to localhost to chat with the built-in WebUI.

You can also query the server as follows.

BASH
curl http://localhost:8080/v1/chat/completions \
    -H "Content-Type: application/json" \
    -d '{
        "messages": [
            {"role": "system", "content": "You are a helpful assistant."},
            {"role": "user", "content": "Write a limerick about python exceptions"}
        ]
    }'

You can also use llama server with coding agents like Pi.

Speculative Decoding

DFlash uses a lightweight block-diffusion drafter model to provide same output with extra speed-ups in decoding phase. Transformers and llama.cpp ship support for DFlash drafter of Muse Glimmer day-0.

Below you can see how speculative decoding can speed-up generation in realistic setups. The video shows llama.cpp webui with DFlash on the left and regular generation on the right.

Speculative Decoding with transformers

You can load the drafter and model as follows, and infer like how you would with base model with an additional parameter (shown in the upcoming snippets).

PYTHON
import torch
from transformers import AutoProcessor, MuseGlimmerAssistantModel, MuseGlimmerForConditionalGeneration

model_id = "meta-models/Muse-Glimmer-30B"
assistant_model_id = "meta-models/Muse-Glimmer-30B-assistant"
target = MuseGlimmerForConditionalGeneration.from_pretrained(model_id, dtype=torch.bfloat16, device_map="auto")
assistant = MuseGlimmerAssistantModel.from_pretrained(assistant_model_id, dtype=torch.bfloat16, device_map="auto")
processor = AutoProcessor.from_pretrained(model_id)

messages = [
    {
        "role": "user", "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/merve/vl-test-suite/resolve/main/SF.png"},
            {"type": "text", "text": "What is shown in this image?"}
        ]
    }
]

inputs = processor.apply_chat_template(
    messages,
    tokenize=True,
    return_dict=True,
    return_tensors="pt",
    add_generation_prompt=True,
    reasoning_strength="low"
).to(target.device)
input_len = inputs["input_ids"].shape[-1]

outputs = target.generate(
    **inputs,
    assistant_model=assistant,
    speculation_type="dflash",
    do_sample=True
)
response = processor.decode(outputs[0][input_len:], skip_special_tokens=False)
print(response)

Speculative Decoding with llama.cpp

You can start llama server using following command. --spec-draft-n-max argument controls how many future tokens DFlash proposes during each speculative-decoding step. Muse Glimmer’s DFlash model was trained with a block size of 16, one anchor token plus 15 proposed tokens, so any value above 15 will be clamped to 15.

BASH
llama serve -hf meta-models/Muse-Glimmer-30B-GGUF --spec-type draft-dflash --spec-draft-n-max 15

You can also use llama cli with speculative decoding drafter as follows.

BASH
llama cli -hf meta-models/Muse-Glimmer-30B-GGUF --spec-type draft-dflash

Inference Endpoints

For a managed, autoscaling deployment, open the Muse Glimmer 30B Inference Endpoints preset. The model is already selected: choose the organization, cloud provider, region, compatible GPU instance, authentication, and autoscaling settings, then review the hourly price and click Create Endpoint. Once its status is Running, you can test it in the Playground and copy the endpoint URL and model name from the Overview.

Deploy Muse Glimmer 30B with Hugging Face Inference Endpoints

The deployed model exposes an OpenAI-compatible Chat Completions API. Keep your Hugging Face token in an environment variable, set HF_ENDPOINT_URL to the URL shown in the Overview without /v1, and set HF_ENDPOINT_MODEL to the endpoint's model name.

BASH
export HF_TOKEN="hf_..."
export HF_ENDPOINT_URL="https://<endpoint-id>.<region>.<cloud>.endpoints.huggingface.cloud"
export HF_ENDPOINT_MODEL="<endpoint-model-name>"
pip install --upgrade openai
PYTHON
import os
from openai import OpenAI

client = OpenAI(
    base_url=f"{os.environ['HF_ENDPOINT_URL'].rstrip('/')}/v1/",
    api_key=os.environ["HF_TOKEN"],
)

response = client.chat.completions.create(
    model=os.environ["HF_ENDPOINT_MODEL"],
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Write a limerick about Python exceptions."},
    ],
    max_tokens=256,
)
print(response.choices[0].message.content)

See the Inference Endpoints documentation for configuration, autoscaling, security, logs, and monitoring.

Support for Muse Glimmer vLLM with transformers backend

For this release, we ship support for vLLM with transformers backend.

BASH

vllm serve meta-models/Muse-Glimmer-30B --model-impl transformers --tensor-parallel-size 4

curl -s http://127.0.0.1:8000/v1/chat/completions \
    -H 'Content-Type: application/json' \
    -d '{
      "model": "meta-models/Muse-Glimmer-30B",
      "messages": [
        {"role": "user", "content": "Explain tensor parallelism briefly."}
      ],
      "temperature": 0.0,
      "max_tokens": 256
    }'

Fine-tuning with TRL

You can use TRL to fine-tune Muse Glimmer using various methods from SFT to Async GRPO. We have run two experiments on bf16 with Hopper-class GPUs with 80GB VRAM each.

WorkloadPractical minimum
Inference / eval, BF161×80 GB H100
LoRA SFT, BF161×80 GB H100, microbatch 1 + checkpointing
Full SFT, BF168×80 GB H100 with FSDP/ZeRO-3
LoRA GRPO, Transformers rollouts1×80 GB H100, but slow/tight
LoRA GRPO, separate vLLM rollout server8×H100: 4 rollout + 4 training
Full-finetune GRPO8 GPUs is usually insufficient

As part of this release, we ship an example to fine-tune Muse Glimmer on small split of MolmoWeb dataset. This shows how to make model generate structured outputs and how to fine-tune on images.

We also experimented with running the model on OpenCode with AsyncGRPO example. Model shows strong coding capabilities, so we encourage you to try training with coding environments in OpenEnv and TRL.

Demos

Here are some fun ways to try out Muse Glimmer. In our opinion, the coolest thing about this model is that it is a local scale personal assistant that can code. That means you can make it do things like, quantize itself, find quantized weights on the Hub, deploy itself to inference endpoints, and even optimize itself for specific hardware! Let’s go team local 🚀

Connect OpenClaw to Muse Glimmer

Assume the Inference Endpoint exposes an OpenAI-compatible /v1 API.

Set HF_TOKEN in the OpenClaw gateway environment, then add this to ~/.openclaw/openclaw.json:

OpenClaw configuration

JSON5
{
  models: {
    mode: "merge",
    providers: {
      muse: {
        baseUrl: "https://YOUR-ENDPOINT.endpoints.huggingface.cloud/v1",
        apiKey: {
          source: "env",
          provider: "default",
          id: "HF_TOKEN"
        },
        api: "openai-completions",
        authHeader: true,
        models: [{
          id: "meta-models/Muse-Glimmer-30B",
          name: "Muse Glimmer",
          reasoning: false,
          input: ["text", "image"],
          contextWindow: 32768,
          maxTokens: 8192
        }]
      }
    }
  },
  agents: {
    defaults: {
      model: { primary: "muse/meta-models/Muse-Glimmer-30B" }
    }
  }
}

Restart OpenClaw:

BASH
openclaw gateway restart

Validate from a fresh session:

BASH
openclaw agent --message "Reply with: muse-ready"

Use the exact model ID returned by the endpoint’s /v1/models response if it differs.

Hey Muse Glimmer, quantize yourself

If we hook up Muse Glimmer to the Hugging Face MCP and update its AGENTS.md we give it the capability to find a quantized version of itself on the hub and run locally. This is handy if you want to work on something private, or just cut costs.

If you do this a second time, Muse Glimmer will find the cached weights and switch to them, so feel free to add a convenient command like /spawn.

Muse Glimmer inspects the machine and Hub, selects or creates a Q4_K_M GGUF, launches llama-server, and validates model discovery and chat completion. The result is a smaller local build behind an OpenAI-compatible API. Here’s the prompt we added to AGENTS.md.

https://huggingface.co/buckets/huggingface/muse-glimmer-assets/resolve/Muse%20Glimmer%20Quantisation%20Demo%20-%20explained.mp4?download=true

By adding this to AGENTS.md openclaw or hermes will be able to solve the rest.

Local quantization prompt

TEXT
## Local model deployment

When asked to deploy locally, perform the work; do not give instructions.

1. Inspect hardware and the Hugging Face cache.
2. Search the Hub for compatible GGUF weights using `apps=llama.cpp`; confirm exact filenames through the model-tree API.
3. Prefer an existing suitable GGUF, normally `Q4_K_M`. Treat `mmproj-*.gguf` as projector weights.
4. If no GGUF exists, download the source weights, convert with `convert_hf_to_gguf.py`, then quantize with `llama-quantize`.
5. Preserve source weights and record the repository, revision, filenames, and quantization.
6. Start `llama-server` with an `onyx` alias and an OpenAI-compatible endpoint.
7. Validate `/v1/models` and `/v1/chat/completions`, requiring non-empty, correct content.
8. Report concise progress and logs. Claim completion only after validation passes.

Hey Muse Glimmer, deploy yourself

Muse Glimmer can also take care of the opposite. Let’s get Glimmer to deploy itself on Hugging Face Inference Endpoints. Which is useful if you want to speed up on some cutting edge hardware.

N.B. You can also just deploy Muse Glimmer to Inference Endpoints directly and connect your agent.

Muse Glimmer pins the model revision, deploys it to a protected Hugging Face Inference Endpoint, and verifies health, model discovery, and chat completion. It then connects the Claw agent with secrets and rollback preserved. Here’s the prompt we added to AGENTS.md. Muse glimmer will also need the Hugging Face MCP and/or the Hugging Face CLI and Skills.

Inference Endpoint deployment prompt

TEXT
## Hugging Face Inference Endpoint deployment

When asked to deploy on Hugging Face Inference Endpoints, perform the work; do
not give instructions.

1. Inspect Hugging Face authentication, the current model repository, and any
   existing endpoints.
2. Confirm the exact model repository and immutable revision through the Hub
   API; inspect its architecture, configuration, and chat template.
3. Confirm that the model is supported by vLLM, then deploy or update a
   protected Inference Endpoint using the managed native vLLM engine.
4. Choose an available region and the smallest suitable accelerator. Use one
   replica and enable scale-to-zero when supported.
5. Preserve the previous endpoint configuration for rollback. Do not expose
   tokens, publish private weights, or replace an unrelated endpoint.
6. Wait for the endpoint to become ready. If startup fails, inspect the logs
   and report the actual blocker rather than repeatedly changing settings.
7. Validate `/health`, `/v1/models`, and `/v1/chat/completions`, requiring the
   expected model and non-empty, correct content. When agent use is required,
   also validate a real structured tool call.
8. Configure the Claw agent to use the endpoint's OpenAI-compatible `/v1` URL,
   storing credentials as secrets and retaining the previous provider as
   rollback. Test the connection in a fresh session.
9. Report concise progress and finish with the repository, revision, engine,
   hardware, endpoint URL, scaling state, and validation results. Claim
   completion only after every required check passes.

Hey Muse Glimmer, optimize yourself

Finally, let’s get Muse Glimmer to do some light RSI. We can instruct our agent to optimize its own inference engine for specific hardware, in this case a Nvidia H100. To do this, the agent will need to use another inference engine, like Inference Endpoints above.

Muse Glimmer benchmarks its own single-H100 serving stack, testing one reversible change at a time while holding the workload fixed. It keeps only correctness-passing gains and finishes with the fastest reproducible configuration. Here’s the prompt we added to AGENTS.md. Muse glimmer need the Hugging Face MCP and the Hugging Face CLI and Skills.

Self-optimization prompt

TEXT
You are Muse Glimmer acting as an autonomous inference-optimization engineer for your own serving stack.

Goal: maximize valid single-H100 aggregate completion throughput in tokens/second.

Protocol:
1. Establish a correctness-passing baseline.
2. Test one reversible optimization at a time.
3. Keep the prompt, concurrency, sampling, request count, warm-up, and decode length fixed.
4. Reject results that fail correctness or prefix checks.
5. Record every experiment chronologically with its configuration, raw throughput, correctness, and delta.
6. Keep improvements and revert regressions.
7. Stop after six consecutive regressions or when the experiment budget is exhausted.
8. Report the best valid configuration and exact reproduction command.

Create a minimal scientific animation of the results:
- white background;
- raw tokens/second—never normalize;
- one point revealed per experiment;
- connect every point chronologically;
- begin with the lowest valid result;
- stop at the best result;
- export as a GIF.

Never fabricate, interpolate, or count correctness-failing measurements.

https://huggingface.co/buckets/huggingface/muse-glimmer-assets/resolve/onyx-optimization-progress.gif?download=true

Hey Muse Glimmer, research the Hub

Try Muse Glimmer as a Hugging Face research agent. The Gradio Space sends each model request to a private Hugging Face Inference Endpoint through its OpenAI-compatible API. It also connects to the official Hugging Face MCP server, giving the agent read-only tools to search and inspect Hub repositories, models, datasets, Spaces, documentation, and papers.

Wrapping Up

We are happy to welcome Muse Glimmer to the Hugging Face Hub. Try Muse Glimmer with your local coding setups today!