Ember-1
Ember-1 is a large language model made by Fireworks, added to the catalog on 24 Sep 2026. It has a context window of 1,048,576 tokens and costs $3.00 per 1M input tokens and $15.00 per 1M output tokens. It accepts text and images and produces text. Typical uses: reasoning over long text, analyzing image-and-text inputs and producing structured responses.
Price checked ยท source: OpenRouter
What Ember-1 is
Fireworks: Ember-1 is a specialized reasoning model from Fireworks Research, built on Kimi K3. It accepts text and image inputs and produces text. Its listed capabilities include reasoning, tool use, structured outputs, and response-format controls, so it can be used for tasks that combine analysis with a defined output format or tool-enabled workflow.
The model has a context window of 1,048,576 tokens and a maximum output of 943,718 tokens. This large context can accommodate extensive source material in one interaction, such as long documents, multiple text passages, or text paired with images. Its text output can present a response to an input or structured content according to the selected response format. The facts specify image input but do not list image generation or other output modalities.
Ember-1 is suited to reasoning work where the prompt includes substantial context, and to workflows that need tool support or structured responses. The combination of text and image input makes it applicable to tasks that require interpreting provided material in either form, while its text-only output keeps results in written form. The available facts do not specify particular tools, supported output schemas, or domain-specific abilities, so those details depend on the surrounding implementation rather than being assumed here.
Key features
- Accepts text and image inputs and produces text
- Supports reasoning and tool use
- Supports structured outputs and response-format controls
- Has a 1,048,576-token context window
- Supports outputs up to 943,718 tokens
Typical uses
- Reasoning over long text
- Analyzing image-and-text inputs
- Producing structured responses
- Tool-supported workflows
Ember-1 pricing
US dollars, list price through OpenRouter.
| Item | Standard |
|---|---|
| Input, per 1M tokens | $3.00 |
| Output, per 1M tokens | $15.00 |
| Cached input read, per 1M tokens | $0.30 |
No price change since it was added on 24 Sep 2026.
| What a task costs | Cost |
|---|---|
| Answer a short question500 in / 300 out | $0.0060 |
| Summarise a 50-page PDF25,000 in / 800 out | $0.087 |
| Review a 2,000-line source file30,000 in / 2,000 out | $0.120 |
| Read a 300-page contract set150,000 in / 1,500 out | $0.473 |
Ember-1 specifications
| Model id | fireworks/ember-1 |
|---|---|
| Version | ember-1-20260923 |
| Context window | 1,048,576 tokensAbout 786,432 English words |
| Max output | 943,718 tokens |
| Knowledge cutoff | Not published |
| Added to catalog | 24 Sep 2026 |
| Input types | text, image |
|---|---|
| Output types | text |
| Tool calling | Yes |
| Structured outputs | Yes |
| Reasoning effort control | Yes |
| Content moderation | Not applied by provider |
Sample outputs
The same prompts are given to every model, and the unedited output is shown with the date it was produced. Upend does not publish leaderboard scores it did not run.
No output has been recorded for Ember-1 yet.