On November 9, 2026, Meta released Muse Ember, which it calls its smallest and most efficient model. It is built for the work agents repeat all day: tool calls, coding subtasks, routing, extraction and long-document parsing.
The short answer: Muse Ember lists at $0.10 per million input tokens and $0.50 per million output tokens — exactly the same as Claude Haiku 5.5 and GPT-6 Luna. So the rate card alone does not tell you which is cheapest. What differs is (1) where each model's long-prompt surcharge starts, (2) the cached-input rate, and (3) how many tokens each model spends to finish a task. Below we line up the official numbers and run three monthly cost scenarios.
What is Muse Ember?
Item | Official detail |
|---|---|
Released | November 9, 2026 (available the same day) |
API model name | muse-ember-1.0 |
Where | Muse Code and Meta Model API; Muse Ember, Muse Spark and Muse Wonder are also listed on Google Cloud and Oracle Cloud |
Context / max output | 1M tokens / 512K tokens |
Inputs | Text, images, audio and video natively; computer use from screenshots |
API | Supports the Responses API |
Training | The first model trained using Muse Wonder Full |
Meta positions its lineup as Muse Wonder for the hardest problems, Muse Spark for coding and agentic workflows, and Muse Ember for repeatable tasks at scale — and suggests pairing them: "Muse Wonder plans, Muse Ember handles the volume."
Muse Ember vs Claude Haiku 5.5 vs GPT-6 Luna pricing
Figures from each vendor's official pricing page (checked November 14, 2026), per 1M tokens:
Muse Ember | Claude Haiku 5.5 | GPT-6 Luna | |
|---|---|---|---|
Input | $0.10 | $0.10 | $0.10 |
Output | $0.50 | $0.50 | $0.50 |
Cached input | $0.004 | $0.01 | $0.01 |
Surcharge starts above | 200K tokens | 100K tokens | 272K tokens |
Above the threshold (input / output) | $0.30 / $1.50 (3x) | $0.50 / $2.50 (5x) | $0.20 / $0.75 |
Context / max output | 1M / 512K | 1M / 128K | 1.05M / 128K |
- Short prompts: the only rate difference is caching, where Muse Ember is 60% cheaper.
- 100K–200K-token prompts: only Haiku 5.5 jumps to 5x, so the gap is widest here.
- Above 272K tokens: GPT-6 Luna's surcharge is the lightest, making it the cheapest of the three.
Monthly cost in three scenarios
Monthly cost = (uncached input × input rate + cached input × cached rate + output × output rate) ÷ 1,000,000 × requests per month. Cache-write fees, reasoning-token differences and tokenizer differences are excluded, so treat these as a guide to how the price structures behave.
Scenario | Per request | Muse Ember | Haiku 5.5 | GPT-6 Luna |
|---|---|---|---|---|
A: FAQ bot, 10,000/month | 20K input (15K cached), 2K output | ~$15.6 | ~$16.5 | ~$16.5 |
B: Contract review, 1,000/month | 150K input, 4K output | ~$17 | ~$85 | ~$17 |
C: Full document set, 1,000/month | 400K input, 8K output | ~$132 | ~$220 | ~$86 |
The winner changes with prompt length. Before switching, bucket your logged requests by input size: under 100K, 100K–200K, 200K–272K and above.
Reading Meta's benchmarks
Benchmark (max reasoning) | Muse Ember | Haiku 5.5 | GPT-6 Luna |
|---|---|---|---|
AutomationBench-AA | 37.8 | 13.9 | 20.2 |
Agents' Last Exam | 61.6 | 57.0 | 57.1 |
Terminal-Bench 4.0 | 40.9 | 39.2 | 16.4 |
FrontierCode 2.0-AA | 27.2 | 34.9 | 25.2 |
OSWorld 2.1 (partial) | 69.5 | 68.9 | 50.5 |
Artificial Analysis Index | 36 | 36 | 28 |
Muse Ember leads clearly on business-workflow automation, but Haiku 5.5 wins on FrontierCode and ties on the Artificial Analysis Index. Meta's core claim is cost per task: Haiku 5.5's best score (max, 36.2) cost $0.411 per task versus $0.136 for Muse Ember (max, 35.8) — about a third. Meta notes these Muse Ember costs are estimates at proposed pricing, so verify token usage on your own workload.
Summary
Muse Ember (muse-ember-1.0) is Meta's smallest Muse model, priced at $0.10 / $0.50 per million tokens — the same as Claude Haiku 5.5 and GPT-6 Luna. It is cheapest on cache-heavy work and in the 100K–200K prompt band, while GPT-6 Luna is cheapest above 272K tokens. Choose by your prompt-length distribution, not the headline rate.
FAQ
How much does Muse Ember cost?
On the Meta Model API, muse-ember-1.0 costs $0.10 per million input tokens, $0.004 per million cached input tokens and $0.50 per million output tokens for prompts up to 200K tokens. Longer prompts cost $0.30, $0.012 and $1.50.
Is Muse Ember cheaper than Claude Haiku 5.5 or GPT-6 Luna?
The headline rates are identical: $0.10 input and $0.50 output per million tokens. The differences are cached input (Muse Ember is $0.004 vs $0.01) and the long-prompt surcharge thresholds: 100K tokens for Haiku 5.5 (5x), 200K for Muse Ember (3x) and 272K for GPT-6 Luna (2x input, 1.5x output).
Where can I use Muse Ember?
Meta says it is available on Muse Code and the Meta Model API as muse-ember-1.0, supports the Responses API, and that Muse Ember, Muse Spark and Muse Wonder are available on Google Cloud and Oracle Cloud.
What are Muse Ember's context window and output limits?
A 1M-token context window and up to 512K output tokens. It natively understands text, images, audio and video.