# Muse Ember Usage, Cost & Rank | OpenCode Data

> Muse Ember ranked #7 by tokens across OpenCode last week, with 0.2% of tokens over the past two months. Muse Ember costs $0.10 per 1M input tokens and $0.50 per 1M output tokens.

Muse Ember is Meta's smallest and most efficient reasoning model, built for the work agents repeat all day: tool calls, coding subtasks, routing, extraction, and long-document parsing. It pairs with Muse Wonder or Muse Spark, which plan while Muse Ember handles the volume.

- Page: https://opencode.ai/data/meta/muse-ember-1-0
- JSON: https://opencode.ai/data/meta/muse-ember-1-0.json
- Model ID: meta/muse-ember-1.0
- Lab: Meta
- Updated: 2026-11-10T05:54:44.000Z

## Model facts

| Fact | Value |
| --- | --- |
| Context window | 1M |
| Max output | 524K |
| Knowledge cutoff | - |
| Release date | 2026-11-09 |
| Input modalities | text, image, video, pdf, audio |
| Output modalities | text |
| Reasoning | Yes |
| Tool calling | Yes |
| Open weights | No |

## Pricing: USD per 1M tokens

| Input | Output | Cached input | Cache write |
| --- | --- | --- | --- |
| $0.10 | $0.50 | $0.0040 | - |

## OpenCode usage: past 2 months

| Metric | Value |
| --- | --- |
| Rank by tokens last week | #7 |
| Tokens | 2.2T |
| Share of all tokens | 0.2% |
| Change vs previous 2 months | +100% |
| Unique users | 41K |
| Completed sessions | 554,695 |
| Average tokens per session | 4M |
| Average cost per session | $0.0000 |
| Total spend | $0 |
| Input tokens served from cache | 0% |
| Weekly retention | - |

## Daily usage: past 2 months

| Date | Tokens | Unique users | Sessions |
| --- | --- | --- | --- |
| 2026-11-03 | 2B | 1.1K | 7,303 |
| 2026-11-04 | 36B | 6K | 54,480 |
| 2026-11-05 | 129B | 5.4K | 47,804 |
| 2026-11-06 | 674B | 7.5K | 79,816 |
| 2026-11-07 | 451B | 6.9K | 95,339 |
| 2026-11-08 | 433B | 6.6K | 126,954 |
| 2026-11-09 | 490B | 7.1K | 142,999 |
| 2026-11-10 | 0 | 0 | 0 |

## Top countries: past 2 months

| Rank | Country | Tokens | Share |
| --- | --- | --- | --- |
| 1 | United States | 334B | 15.1% |
| 2 | India | 254B | 11.4% |
| 3 | Brazil | 153B | 6.9% |
| 4 | Indonesia | 120B | 5.4% |
| 5 | Egypt | 102B | 4.6% |
| 6 | Vietnam | 74B | 3.3% |
| 7 | Türkiye | 72B | 3.2% |
| 8 | United Kingdom | 66B | 3% |
| 9 | Canada | 59B | 2.7% |
| 10 | Japan | 55B | 2.5% |
| 11 | Ukraine | 49B | 2.2% |
| 12 | Bangladesh | 48B | 2.1% |
| 13 | Philippines | 47B | 2.1% |
| 14 | South Korea | 41B | 1.8% |
| 15 | Algeria | 39B | 1.7% |

## Nearby models by tokens last week

| Rank | Model | Lab | Tokens |
| --- | --- | --- | --- |
| 3 | [muse-spark-1.3-contributor](https://opencode.ai/data/meta/muse-spark-1-3-contributor.md) | Meta | 6T |
| 4 | [beam-1.0-501b-a23b](https://opencode.ai/data/reflection/beam-1-0-501b-a23b.md) | Reflection | 5T |
| 5 | [mimo-v2.6-flash](https://opencode.ai/data/xiaomi/mimo-v2-6-flash.md) | Xiaomi | 3.2T |
| 6 | [nemotron-3.5-super-vl](https://opencode.ai/data/nvidia/nemotron-3-5-super-vl.md) | NVIDIA | 3.1T |
| 7 | [muse-ember-1.0](https://opencode.ai/data/meta/muse-ember-1-0.md) | Meta | 2.2T |
| 8 | [inkling-2-small](https://opencode.ai/data/thinkingmachines/inkling-2-small.md) | Thinking Machines | 1.7T |
| 9 | [pulse-alpha](https://opencode.ai/data/unknown/pulse-alpha.md) | - | 1.2T |
| 10 | [gpt-6-luna](https://opencode.ai/data/openai/gpt-6-luna.md) | OpenAI | 979B |
| 11 | [deepseek-v4.1-flash-1017](https://opencode.ai/data/deepseek/deepseek-v4-1-flash-1017.md) | DeepSeek | 919B |
| 12 | [deepseek-v4-pro](https://opencode.ai/data/deepseek/deepseek-v4-pro.md) | DeepSeek | 870B |

## Benchmarks

| Benchmark | Score | Metric | Source |
| --- | --- | --- | --- |
| AutomationBench | 37.8 | strict pass@1 | https://research.meta.ai/blog/muse-ember |
| Agents' Last Exam | 61.6 | score | https://research.meta.ai/blog/muse-ember |
| Terminal-Bench | 40.9 | accuracy | https://research.meta.ai/blog/muse-ember |
| FrontierCode | 27.2 | score | https://research.meta.ai/blog/muse-ember |
| GDP.pdf | 22.8 | score | https://research.meta.ai/blog/muse-ember |
| OSWorld | 69.5 | mean partial score | https://research.meta.ai/blog/muse-ember |
| OSWorld | 33.8 | binary completion rate | https://research.meta.ai/blog/muse-ember |
| PerceptionBench | 80.6 | score | https://research.meta.ai/blog/muse-ember |
| Artificial Analysis Intelligence Index | 36 | index score | https://research.meta.ai/blog/muse-ember |

## Methodology

- Updates: Aggregated every hour. Days and weeks use UTC.
- Tokens: Input, output, reasoning, and cached tokens for each request.
- Users and sessions: Approximate counts of distinct users and OpenCode sessions.
- Cost: Session cost is the average cost per OpenCode session. Token prices are list prices from the OpenCode model catalog.
- Retention: The share of a model's users in one week who use it again the next week.
- Citation: Cite OpenCode Data (opencode.ai/data) with the update time shown at the top of the page.