Meta is back with Muse Glimmer: local, agentic, multimodal, and open source

Meta released Muse Glimmer, a 30B-parameter multimodal AI model under the Apache 2.0 license, designed for local deployment in privacy-aware applications such as coding and document analysis. The model features a 2B vision encoder and a 28B text decoder with hybrid attention, and Hugging Face announced day-0 support in transformers, llama.cpp, vLLM, and Inference Endpoints.

Image-Text-to-Text • Updated • 194 Meta is back with Muse Glimmer: local, agentic, multimodal, and open source Update on GitHub https://github.com/huggingface/blog/blob/main/muse-glimmer.md 30B parameters, and released under the Apache 2.0 license , it’s ideal deploying locally for privacy, reducing costs, or just hacking around. It’s intended for privacy-aware applications such as coding, document analysis, personal assistants, Claw- or Hermes-like setups. To celebrate, we are shipping with Meta day-0 support in transformers , llama.cpp , vLLM , Inference Endpoints, and other libraries. We built a few cool things and explain our findings in this blog. Check out the demos below for inspiration. You can find all Muse Glimmer models in this collection https://huggingface.co/collections/meta-models/muse-glimmer . Architecture Muse Glimmer is a dense 30B parameter model consisting of: - 2B ViT-style encoder for vision Perception Encoder - 28B parameter text decoder In addition to the main VLM, there’s also a speculative decoding drafter implemented on DFlash. Usage of this module is optional, and it can provide much faster generation in exchange for some memory cost. We found this drafter to be particularly well suited to structured content generation such as coding. Text Decoder The language model uses the following architecture components: 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 https://huggingface.co/papers/2504.13181 . 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. 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/Muse-Glimmer-30B" Load model 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. Prompt messages = {"role": "user", "content": "Write a short joke about saving RAM."}, Process input 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 Generate output outputs = model.generate inputs, max new tokens=1024 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. pip install torchvision Muse Glimmer accepts images as input, as demonstrated here: 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 Generate output outputs = model.generate inputs, max new tokens=512 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 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, max new tokens=128 response = processor.decode outputs 0 input len: , skip special tokens=False parsed = processor.tokenizer.parse response response Object Detection You can use Muse Glimmer to do open ended object detection in images as follows. python import json 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, .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 {"x min": 0, "y min": 390, "x max": 520, "y max": 603} note that you need to scale X and Y values to image size to visualize: xyxy = round box "x min" / 1000 width , round box "y min" / 1000 height , round box "x max" / 1000 width , round box "y max" / 1000 height , Video Inference To work with videos we recommend installing torchcodec into the environment. 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. messages = {"role": "system", "content": "You are a helpful assistant."}, { "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, max new tokens=1024 response = processor.decode outputs 0, input len: , skip special tokens=False, parsed = processor.parse response response, prefix=inputs "input ids" , print parsed Llama.cpp Muse Glimmer comes with day-0 llama.cpp support. Meta has distributed calibrated quants in this repo https://huggingface.co/meta-models/Muse-Glimmer-30B-GGUF , and Uunsloth 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 curl -LsSf https://llama.app/install.sh | sh Then you can start the server as follows. llama serve meta-models/Muse-Glimmer-30B-GGUF Once the server has started, you can head to localhost:8080 to chat with the built-in WebUI. TODO: Insert webui video with this model You can also query the server as follows. 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 transformersYou 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" target = MuseGlimmerForConditionalGeneration.from pretrained model id, dtype=torch.bfloat16, device map="auto" assistant = MuseGlimmerAssistantModel.from pretrained 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?"} } out = target.generate inputs, assistant model=assistant, speculation type="dflash", max new tokens=64, do sample=True print processor.batch decode out 0 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. 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. llama cli -hf meta-models/Muse-Glimmer-30B-GGUF --spec-type draft-dflash Support for Muse Glimmer vLLM with transformers backend For this release, we ship support for vLLM with transformers backend. tensor parallel serving across 4 GPUs vllm serve meta-models/Muse-Glimmer-30B --model-impl transformers --tensor-parallel-size 4 infer curl -s http://127.0.0.1:8000/v1/chat/completions \ -H 'Content-Type: application/json' \ -d '{ "model": "username/muse-glimmer-hf-v2", "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. | Workload | Practical minimum | |---|---| | Inference / eval, BF16 | 1×80 GB H100 | | LoRA SFT, BF16 | 1×80 GB H100, microbatch 1 + checkpointing | | Full SFT, BF16 | 8×80 GB H100 with FSDP/ZeRO-3 | | LoRA GRPO, Transformers rollouts | 1×80 GB H100, but slow/tight | | LoRA GRPO, separate vLLM rollout server | 8×H100: 4 rollout + 4 training | | Full-finetune GRPO | 8 GPUs is usually insufficient | As part of this release, we ship an example to fine-tune Muse Glimmer on small split of MolmoWeb dataset https://huggingface.co/merve/smol-vision/blob/main/qlora click grounding.ipynb . 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 https://github.com/huggingface/trl/blob/main/examples/scripts/openenv/opencode.py . Model shows strong coding capabilities, so we encourage you to try training with coding environments. 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 { 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/Muse-Glimmer-30B", name: "Muse Glimmer", reasoning: false, input: "text", "image" , contextWindow: 32768, maxTokens: 8192 } } } }, agents: { defaults: { model: { primary: "muse/meta/Muse-Glimmer-30B" } } } } Restart OpenClaw: openclaw gateway restart Validate from a fresh session: 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 https://huggingface.co/mcp and update its AGENTS.md http://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 . By adding this to AGENTS.md http://AGENTS.md openclaw or hermes will be able to solve the rest. Local quantization prompt 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 https://endpoints.huggingface.co/huggingface/new/meta-models/Muse-Glimmer-30B 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 http://AGENTS.md . Muse glimmer will also need the Hugging Face MCP https://huggingface.co/mcp and/or the Hugging Face CLI and Skills https://huggingface.co/docs/hub/en/agents-skills . Inference Endpoint deployment prompt 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 http://AGENTS.md . Muse glimmer need the Hugging Face MCP https://huggingface.co/mcp and the Hugging Face CLI and Skills https://huggingface.co/docs/hub/en/agents-skills . Self-optimization prompt 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. Wrapping Up We are happy to welcome Muse Glimmer in Hugging Face Hub. Try the models in this collection https://huggingface.co/collections/meta-models/muse-glimmer with your local coding setups today 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.