Skip to content
All systems operational0 AI providers monitored, polled every 2 minutes
Live status

Muse Ember

Budget

by Meta

As of November 16, 2026, Muse Ember by Meta costs $0.10 per 1M input tokens and $0.50 per 1M output tokens, with a 1M-token context window.

Muse Ember is the smallest model Meta Superintelligence Labs sells, shipped on November 9, 2026 as muse-ember-1.0 and pitched squarely at price per task rather than capability. Meta describes it as built for high-volume agentic workflow execution, which is the honest framing: it is the executor you put underneath a planner, not the planner. The rate card has a cliff rather than a flat line. Prompts up to 200,000 input tokens bill at $0.10 per million input tokens, $0.004 on a cached read, and $0.50 output; cross 200,000 and the same request bills at $0.30, $0.012, and $1.50, three times as much on every line. The $0.10 and $0.50 floor is exactly the Claude Haiku 5.5 short-prompt rate and the GPT-6 Luna rate, and the cached read is the cheapest of the three by a wide margin. The context window is 1 million tokens, input covers text, image, audio, and video, and reasoning effort is configurable from minimal through max. It is available on the Meta Model API and included in all Muse Code subscription plans, and Meta has not published weights for it. Meta's launch table is vendor-reported but unusually specific about where the model loses, which makes it worth reading line by line: an Artificial Analysis Intelligence Index of 36 at max effort, level with Claude Haiku 5.5 at 36 and well ahead of GPT-6 Luna at 28, at an estimated $0.13 per Index task; AutomationBench-AA at 37.8 against 13.9 for Haiku 5.5 and 20.2 for Luna; Agents' Last Exam at 61.6 against 57.0 and 57.1; a Personal Superintelligence Index of 46.8 against 36.1 and 37.0; Terminal-Bench 4.0 at 40.9 against 39.2 and 16.4; a Coding Agent Index-AA of 46 against 37 and 31; OSWorld 2.1 at 69.5 on the partial setting and 33.8 binary; and PerceptionBench with tools at 80.6. The exception is the one that matters most for coding: FrontierCode 2.0-AA comes in at 27.2 against 34.9 for Haiku 5.5, so Ember trails the model it otherwise matches on exactly the work most teams would buy it for. No independent evaluator has scored it yet.

Input Price

$0.10

per 1M tokens

Output Price

$0.50

per 1M tokens

Context Window

1M

tokens

Released

2026-11

API access

Capabilities

textvisionaudiovideotool-usecodereasoning

Key Strengths

  • ✓$0.10 / $0.50 per 1M tokens up to a 200K input prompt, tripling above it
  • ✓Cached reads at $0.004, below Claude Haiku 5.5 and GPT-6 Luna
  • ✓1M token context window
  • ✓Text, image, audio, and video input in the cheapest Meta model
  • ✓Artificial Analysis Intelligence Index of 36, level with Claude Haiku 5.5
  • ✓Included in every Muse Code subscription plan
  • ✓Configurable reasoning effort from minimal through max

Best For

  • ▸High-volume agent steps under a more capable planner
  • ▸Classification, extraction, and routing at scale
  • ▸Multimodal pipelines over images, audio, and video
  • ▸Cost-sensitive computer use and automation work

Pricing Details

Input tokens

$0.10

per 1M tokens

Output tokens

$0.50

per 1M tokens

Estimated cost per 1K requests

$0.35

~1K input + ~500 output tokens avg

Prices are subject to change. Check the official documentation for current pricing. See the cost calculator for detailed estimates.

Quick answers

How much does Muse Ember cost?
As of November 16, 2026, Muse Ember costs $0.10 per 1M input tokens and $0.50 per 1M output tokens on the Meta API.
What is the context window of Muse Ember?
Muse Ember has a 1M-token context window (1,000,000 tokens).

Related Models

View DocumentationCompare ModelsCost CalculatorFull Pricing Guide