Muse Glimmer 30B
Neural Network
Muse Glimmer 30B is an open multimodal model from Meta Superintelligence Labs for autonomous agents and long-horizon tasks.
Specifications
- Input
- text, images, documents
- Output
- text
- Context size
- 131 072 tokens
- Max answer length
- 16 384 tokens
- Reasoning effort
- low, medium, high, extra high
- Capabilities
- reasoning, function calling
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works Muse Glimmer 30B?
Muse Glimmer 30B is a dense open-weights model from Meta Superintelligence Labs, distilled from Muse Spark and optimized for consumer hardware. It is multimodal and focused on autonomous agents and long-horizon tasks where the result consists of many sequential steps. Via BotHub, the model is accessible from Russia without a VPN or foreign card, with pay-as-you-go billing in rubles for tokens, no subscription, and no balance expiration. Other neural networks are available in the same window: compare answers and switch without rewriting integrations using the unified OpenAI-compatible API. Data is transmitted with AES-GCM encryption, and companies have access to contracts, invoices, electronic document management, and an admin panel with employee limits. Use cases include: agents that independently execute long action chains; coding and technical documentation assistants; analyzing materials with mixed text and images; and debugging scenarios before deploying to your own infrastructure with open weights.Frequently asked questions about Muse Glimmer 30B
You can use generated results for commercial purposes. You own all rights to the content you create. The only restriction: make sure your prompt does not include copyrighted third-party material. You are responsible for respecting the rights to any input data.
muse-glimmer-30b is a text model from Meta: it writes and edits text, analyzes code, and answers based on documents and images. A context window of up to 131,072 tokens allows loading large materials in full, and the reasoning mode helps where step-by-step task analysis is needed.
In a single response, the model outputs up to 117,964 tokens — enough for a voluminous document, a large code module, or a long analysis without splitting into parts. Payment on BotHub is pay-as-you-go for tokens, no subscription is needed, and the balance does not expire.
Yes, the model has a reasoning mode: before the final answer, it breaks down the task step-by-step. This is useful in logic, mathematics, requirements analysis, and code debugging — where not only the result but also a clear solution path is important.
Yes, the model supports function calling and structured JSON output, making it easy to integrate into agents and services where the response needs to be parsed programmatically. On BotHub, it is available via a unified OpenAI-compatible API — you can switch models without rewriting the integration.
You can input images and documents along with text: the model will analyze a screenshot, diagram, table, or file and answer based on the content. There is no note in our data regarding audio and video input, so please check the required format with your own example.
There is no specific note about languages in our data, so it is better to check the quality of Russian on your own task. Open the model on the website or in the Telegram bot, provide your text, and compare the answer with another model in the same window. Access from Russia is available without a VPN or foreign card.