Primary use stated or inferred cautiously from official documentation.
Muse Ember
Meta multimodal model released on November 9 for frequent coding and analysis tasks. It understands text, images, audio and video and returns text responses.
The essential
Maximum input capacity when published by the source.
Documented maximum generation limit.
Verified availability channels.
Primary functional family and verified specialties.
Declared terms for API access, model weights, or self-hosting.
Limits and integration
| API ID | muse-ember-1.0 |
|---|---|
| Model type | Text and reasoning · Code · Multimodal |
| Access model | Paid proprietary |
| License | Proprietary |
| Deployment | Hosted API |
| Release | 09 NOV 2026 |
| Knowledge cutoff | Not published |
| Entrance | Text · Image · Audio · Video |
| Exit | Text · Code |
| Context window | 1,000,000 tokens |
| maximum output | 512,000 tokens |
| Reasoning | Not published |
| Published tools | |
| Structured Outings | Not published |
| Batch processing | Not published |
| Prompt cache | Not published |
| Fine-tuning | Not published |
| Verified platforms | Meta Model API · Muse Code · Google Cloud · Oracle Cloud |
Review long documents, interpret screenshots and charts or solve coding subtasks. Meta serves it through the Responses API with a one-million-token window.
Limits apply to Meta’s API. Other providers may document different limits and modalities. A cost comparison also needs to account for output tokens and prompt length.
What it can do
Orientation
Review long documents, interpret screenshots and charts or solve coding subtasks. Meta serves it through the Responses API with a one-million-token window.
Context
1,000,000 tokens · 512,000 tokens
Tools and integration
Not published
Access
Meta Model API · Muse Code
Documented cost
| Concept | Worth | Unit/condition |
|---|---|---|
| Standard input | $0.10 / 1M tokens | Standard API rate |
| Cached input | Not published | Reading reused prefixes |
| Cache write or storage | Not published | The condition varies by provider |
| Standard output | $0.50 / 1M tokens | May include reasoning tokens |
| Batch input | Not published | Asynchronous processing |
| Batch output | Not published | Asynchronous processing |
For prompts up to 200,000 tokens, input costs $0.10, cached input $0.004 and output $0.50 per million. Above that threshold the rates are $0.30, $0.012 and $1.50 respectively.
i Prices change and may depend on level, region or context length. Check the source before making a decision.
How to read the results
Meta reports 37.8% on AutomationBench-AA and 40.9% on Terminal-Bench 4.0. The first figure is strict pass@1. The second used mini-swe-agent, so the harness affects comparisons.
| Benchmark | Result | Metric | Source |
|---|---|---|---|
| AutomationBench-AA | 37,8 % | pass@1 | View source ↗ |
| Terminal-Bench 4.0 | 40,9 % | Accuracy | View source ↗ |
Inferama only highlights a “best result” when the metric, test set, configuration, and date allow for an equivalent comparison. The supplier's figures are presented as claims from its own source.
Chronology
Muse Ember 1.0
Generally available via API