Muse Glimmer 30B
Muse Glimmer 30B is Meta's Apache 2.0 licensed 30B dense model for agentic work, accepting interleaved text and images across 100+ languages within a context window of 131.1K tokens.
View API reference- Input and output price
- Prices from: Input $0.30, Output $1.10, Per 1M tokens
- 24h uptime
- Loading AI Gateway uptime
import { streamText } from 'ai'
const result = streamText({ model: 'meta/muse-glimmer-30b', prompt: 'Why is the sky blue?'})Copy link to headingPlayground
Try out Muse Glimmer 30B by Meta. Usage is billed to your team at API rates. Free users (those who haven't made a payment) get $5 of credits every 30 days.
Muse Glimmer 30B
Muse Glimmer 30B
Copy link to headingProviders
Route requests across multiple providers. Copy a provider slug to set your preference. Visit the docs for more info. Using a provider means you agree to their terms, listed under Legal.
Copy link to headingUptime24 hours
Direct request success rate on AI Gateway and per-provider. Visit the docs for more info.
Copy link to headingThroughput24 hours
P50 throughput on live AI Gateway traffic, in tokens per second (TPS). Visit the docs for more info.
Copy link to headingLatency24 hours
P50 time to first token (TTFT) on live AI Gateway traffic, in milliseconds. View the docs for more info.
Getting started
Call Muse Glimmer 30B through AI Gateway with the AI SDK generateText and streamText functions, or through the OpenAI Chat Completions, OpenAI Responses, and Anthropic Messages APIs by changing the base URL. AI Gateway authenticates the request and routes it to an available provider.
Install the AI SDK (pnpm add ai dotenv), create an API key from the API Keys page, and set it as AI_GATEWAY_API_KEY in your environment. Full setup is covered in the text generation quickstart.
import { generateText } from 'ai';import 'dotenv/config';
async function main() { const result = await generateText({ model: 'meta/muse-glimmer-30b', prompt: 'Why is the sky blue?', });
console.log(result.text);}
main().catch(console.error);Top-level parameters
The same Muse Glimmer 30B request in each API format AI Gateway supports.
import { generateText } from 'ai';import 'dotenv/config';
async function main() { const result = await generateText({ model: 'meta/muse-glimmer-30b', system: 'You are a concise technical assistant.', prompt: 'Summarize the tradeoffs between static generation and SSR.', maxOutputTokens: 1024, });
console.log(result.text);}
main().catch(console.error);Standard parameters like prompt, messages, temperature, and tools work as documented in the AI SDK docs. These are the parameters with model-specific behavior.
| Parameter | Type | Required | Description |
|---|---|---|---|
model | string | Yes | Model ID in the form creator/model, e.g. meta/muse-glimmer-30b. AI Gateway routes the request to an available provider. |
maxOutputTokens | number | No | Hard cap on generated tokens. Muse Glimmer 30B supports up to 131,072 output tokens. Reasoning tokens count toward this limit. |
reasoning | 'provider-default' | 'none' | 'minimal' | 'low' | 'medium' | 'high' | 'xhigh' | No | Provider-agnostic reasoning effort, available in AI SDK 7 or later. Maps to the provider’s native reasoning configuration; reasoning settings under providerOptions take precedence when both are set. See the Reasoning section below. |
providerOptions | Record<string, JSONValue> | No | AI Gateway routing options under gateway, plus any provider-native options under the provider’s own namespace — see the table below. |
Input limits
| Input | Formats | Sources | Max count | Max size | Limits |
|---|---|---|---|---|---|
| Text | — | — | — | — | Prompt and response share the 131K-token context window |
| Image | — | URL, base64, Uint8Array | — | — | Sent as image parts in messages; counts as input tokens |
Provider options
Set AI Gateway routing options under providerOptions.gateway. For provider-specific options, pass them under the provider’s namespace as documented by the AI SDK.
Learn more in the AI SDK provider docs.
import { generateText } from 'ai';import 'dotenv/config';
async function main() { const result = await generateText({ model: 'meta/muse-glimmer-30b', prompt: 'Why is the sky blue?', providerOptions: { gateway: { only: ['togetherai', 'parasail'], }, }, });
console.log(result.text);}
main().catch(console.error);These AI Gateway routing options apply to every model. Provider-specific options pass through under the provider’s own namespace (for example providerOptions.anthropic) exactly as documented by the AI SDK.
| Parameter | Type | Required | Description |
|---|---|---|---|
providerOptions.gateway.only | string[] | No | Restrict routing to these provider slugs. Requests fail over only within the listed providers. |
providerOptions.gateway.order | string[] | No | Preferred provider order. Listed providers are tried first; unlisted providers remain available as fallbacks. |
providerOptions.gateway.sort | 'cost' | 'ttft' | 'tps' | No | Rank candidate providers by price, time to first token, or tokens per second instead of the default routing order. |
providerOptions.gateway.zeroDataRetention | boolean | No | Route only to providers with a zero-data-retention policy for this model. |
Routing across providers
AI Gateway serves the same model through multiple providers and fails over automatically. order expresses a preference while keeping every provider eligible; only is a hard allowlist — if none of the listed providers are available the request fails instead of falling back.
Options under a provider's own namespace (for example providerOptions.anthropic) are forwarded to that provider with the request. Providers ignore option namespaces that don't apply to them, so it is safe to set provider options alongside gateway routing options.
Reasoning
AI Gateway bridges reasoning across every API format. The AI SDK exposes a provider-agnostic top-level reasoning level (none, minimal, low, medium, high, or xhigh); the Chat Completions and Responses formats take the same effort under reasoning.effort; and the Anthropic Messages format uses a native thinking token budget. Whichever you send, the gateway maps it to the target model’s native configuration, converting between effort levels and token budgets as needed. Reasoning-related settings under providerOptions take full precedence over the top-level reasoning value and are never merged. Reasoning tokens typically count toward your output-token usage, though how they’re reported and billed varies by provider.
Learn more in the AI Gateway reasoning guide.
import { generateText } from 'ai';import 'dotenv/config';
async function main() { const result = await generateText({ model: 'meta/muse-glimmer-30b', prompt: 'Explain the Monty Hall problem step by step.', reasoning: 'high', });
console.log(result.text);}
main().catch(console.error);Image input
Send images alongside text as message parts. Images count as input tokens.
import { generateText } from 'ai';import 'dotenv/config';
async function main() { const result = await generateText({ model: 'meta/muse-glimmer-30b', messages: [ { role: 'user', content: [ { type: 'text', text: 'Describe this image.' }, { type: 'image', image: 'https://example.com/photo.jpg' }, ], }, ], });
console.log(result.text);}
main().catch(console.error);Tool calling
Expose tools the model can call. Define each tool’s inputs with a Zod schema.
import { generateText, tool } from 'ai';import { z } from 'zod';import 'dotenv/config';
async function main() { const result = await generateText({ model: 'meta/muse-glimmer-30b', prompt: 'What is the weather in San Francisco?', tools: { getWeather: tool({ description: 'Get the current weather for a location', inputSchema: z.object({ location: z.string() }), execute: async ({ location }) => ({ location, temperatureC: 18 }), }), }, });
console.log(result.text);}
main().catch(console.error);Copy link to headingMore models by Meta
| Model |
|---|
| — | $0.01/img | — | | 08/26/2026 | ||||||||||
| 1M | 5.1 s | 130 tps | $0.10/M | $0.20/M | Read$0.002/M | — | +2 | | 08/05/2026 | |||||
| 1M | 4.0 s | 149 tps | $1.25/M | $4.25/M | Read$0.15/M | — | +2 | | 08/05/2026 | |||||
| 1M | 2.8 s | 169 tps | $1.25/M | $4.25/M | Read$0.15/M | — | +2 | | 07/09/2026 | |||||
| 1M | 4.1 s | 91 tps | $0.10/M | $0.20/M | Read$0.002/M | — | +2 | | — | |||||
| 1M | 4.7 s | 94 tps | $1.25/M | $4.25/M | Read$0.15/M | — | +2 | | — |
Copy link to headingAbout Muse Glimmer 30B
Muse Glimmer 30B was released August 10, 2026 as Meta's open-weight model for agentic work, licensed under Apache 2.0. It is a dense causal transformer of roughly 29.6 billion parameters across 52 layers, including a perception encoder that lets it accept interleaved text and images, and it supports more than 100 languages within a context window of 131.1K tokens.
The attention design matters for how it behaves at length: grouped-query attention with a repeating local-local-local-global pattern and a sliding window, so most layers attend locally and periodic global layers carry longer-range structure. Reasoning effort is selectable across low, medium, high, and xhigh through the system prompt.
Muse Glimmer 30B is a distillation of Meta's flagship, which shapes where it is strong. On agentic orchestration and reasoning it leads comparable open models, with 75.5 on MCP Atlas, 74.6 on DeepSearch QA, and 94.7 on AIME 2026. It trails on computer-use and terminal work, where Qwen3.6-27B stays ahead on OSWorld-Verified and TerminalBench 2.1. The knowledge cutoff is January 4, 2026.
Quantized to roughly 4 bits the model drops under 20 GB, which leaves room on a 24 to 32 GB machine for the KV cache and encoder. Through AI Gateway you can call it over an API instead, and keep the option of self-hosting later.
You can integrate Muse Glimmer 30B through AI SDK, Chat Completions API, Responses API, Messages API, or other API formats, from TypeScript or Python.
Copy link to headingWhat To Consider When Choosing a Provider
- Configuration: Muse Glimmer 30B is deliberately less capable than the Muse Spark flagship it was distilled from, and Meta says it does not meet the Frontier AI definition in its own scaling framework. Treat it as a capable open model rather than a flagship substitute.
- Configuration: Strength is uneven by task type. Agentic orchestration and reasoning lead the comparable open models, while computer-use and terminal work trail them. Match it to the first kind of workload rather than assuming the lead generalises.
- Configuration: On CI Memories, a privacy benchmark where a lower score is better, Muse Glimmer 30B scores worse than Gemma and better than Qwen, so it is not uniformly stronger than its peers. Meta recommends deploying it with guardrails, including human confirmation before irreversible actions.
- Configuration: The knowledge cutoff is January 4, 2026, so pair it with web search or retrieval for anything more recent.
- Zero Data Retention: Zero Data Retention is available for this model. It is offered on a per-provider and model basis. See the documentation for details.
- Authentication: AI Gateway authenticates requests using an API key or OIDC token. You do not need to manage provider credentials directly.
Copy link to headingWhen to Use Muse Glimmer 30B
Best for
- Agentic Orchestration: Leading comparable open models on MCP Atlas
- Apache 2.0 Licensing: No revenue threshold or commercial gate
- Local Deployment: Roughly 4-bit on a 24 to 32 GB machine
- Multilingual Work: More than 100 languages supported
- Interleaved Text And Images: Handled by a built-in perception encoder
Consider alternatives when
- Frontier Capability: Muse Spark 1.2 is the flagship it was distilled from
- Computer-Use And Terminal Work: Comparable open models score higher
- Privacy Benchmark Selection: Its CI Memories result is mid-pack
- Post-Cutoff Facts: January 2026 knowledge needs web search or retrieval
Copy link to headingConclusion
Muse Glimmer 30B is an Apache 2.0 licensed 30B model that leads comparable open models on agentic orchestration and runs on a single consumer GPU when quantized. Point meta/muse-glimmer-30b at AI Gateway to try it without local hardware, and deploy it with guardrails on irreversible actions as Meta recommends.
Copy link to headingFrequently Asked Questions
What licence does Muse Glimmer 30B use?
Apache 2.0, with no revenue threshold or separate commercial licence required.
How large is Muse Glimmer 30B?
Roughly 29.6 billion parameters in a dense transformer across 52 layers, including a perception encoder for image input.
Can I run Muse Glimmer 30B locally?
Yes. At roughly 4-bit quantization it drops under 20 GB, leaving room on a 24 to 32 GB machine for the KV cache and encoder. Through AI Gateway you can call it over an API instead.
What is Muse Glimmer 30B strongest at?
Agentic orchestration and reasoning. It reaches 75.5 on MCP Atlas, 74.6 on DeepSearch QA, and 94.7 on AIME 2026, ahead of comparable open models. It trails them on computer-use and terminal work.
How does Muse Glimmer 30B relate to Muse Spark?
It is a distillation of the flagship. Meta states it does not meet the Frontier AI definition in its own scaling framework, so treat it as a capable open model rather than a flagship substitute.
What is the context window for Muse Glimmer 30B?
The context window is 131.1K tokens, with up to 131.1K tokens per response.
What safety guidance does Meta give?
Deploy with guardrails, including human confirmation before irreversible actions. On CI Memories, a privacy benchmark where lower is better, it scores mid-pack against comparable open models.
Does Muse Glimmer 30B support Zero Data Retention?
Yes, Zero Data Retention is available for this model. Zero Data Retention is offered on a per-provider basis. See https://vercel.com/docs/ai-gateway/capabilities/zdr for details.