The AI arena is free today

Open Superagent
FireworksReleased on Sep 23, 2026

Ember-1: Benchmarks, Pricing & Context Window

Ember-1 is a language model from Fireworks, released in September 2026, with a 1.0M-token context window, and pricing from $3.00/M input, $0.300/M cached input, $15.00/M output.

Ember-1 is a Fireworks Research specialized model built on Moonshot Kimi K3 that shortens reasoning traces (~40% fewer tokens) while matching K3-max quality on Fireworks' evaluations. Research Preview on Fireworks Serverless (model path

Input
Text
Output
Text

Ember-1 benchmarks

Capability tiers

Standing within each category, adjusted for leaderboard depth.

Real tasks performance

High-confidence performance for Ember-1 across real-world prompt categories. Only 95% intervals at most 4 points wide are shown.

Performance by conversation depth

How Ember-1 holds up as conversations get longer.

Quality Tracker

Ember-1 Performance Across Datasets

Scores sourced from the model's scorecard, paper, or official blog posts

LLM Stats Logollm-stats.com - Wed Oct 07 2026
Notice missing or incorrect data?

Ember-1 pricing

Providers

Ember-1 starts at $3.00 per million input tokens and $15.00 per million output tokens via Fireworks. Reused prompt prefixes cost $0.300 per million cached input tokens.

ProviderInput $/MCached input $/MOutput $/MContext in / outTTFT p95 sOutput p5 c/sModalities in / out
Fireworks logoFireworks
$3.00$0.300$15.001.0M/—
—
—
/

Cached input is the discounted price for prompt tokens served from a provider cache. TTFT is time to first token. Output is characters per second; p5 is the sustained floor exceeded by 95% of observed requests.

Ember-1 context window

Input and output token limits for Ember-1, plus how it ranks on long-context understanding.

Input
1.0Mtokens
≈ 1.6k pages of text
1.0M
8K128K1M

Try now

huggle
Ember-1in Huggle

Make it with
Ember-1.

Ember-1

Ember-1 latency

Ember-1 time to first token, sustained output throughput, and failed-request rate from live model usage over the trailing 7 days.

Ember-1 examples

Recent arena outputs from Ember-1, picked from the highest-ranked matchups.

Ember-1 license

Ember-1 is a proprietary model available under its provider's product and API terms, has 2.8T parameters.

License
Proprietary
Hosted access
Parameters
2.8T

Proprietary license - usage restrictions apply

Ember-1 resources

Official sources for Ember-1: provider documentation, official launch post.

Ember-1 vs other models

The most-compared alternatives to Ember-1 are Claude Mythos Preview, GPT-5.1 Thinking, GPT-5.1 Instant. Open any pair side-by-side for benchmarks, pricing, context, and latency.

Models like Ember-1

Models ranked just above and below Ember-1 by LLM Stats score.

 

Claude Mythos Preview

Score pending
 

GPT-5.1 Thinking

Score pending
 

GPT-5.1 Instant

Score pending
 

Nova 2 Pro

Score pending
 

Muse Spark 1.3

Score pending
 

Atria Dawn Preview

Score pending

FAQ

Common questions about Ember-1.

When was Ember-1 released?

Ember-1 was released on September 23, 2026 by Fireworks. This is the official Ember-1 release date tracked on LLM Stats.

How much does Ember-1 cost?

Ember-1 pricing starts at $3.00 per million input tokens, $0.30 per million cached input tokens, $15.00 per million output tokens via Fireworks, the lowest price among tracked providers.

How big is Ember-1?

Ember-1 has 2780 billion parameters.

Who created Ember-1?

Ember-1 was created by Fireworks.

What is the license for Ember-1?

Ember-1 is released under the Proprietary license.

Where can I use Ember-1?

Ember-1 is available through 1 provider including Fireworks.

Where is the Ember-1 paper or technical report?

Ember-1 has a paper or technical report available at https://fireworks.ai/blog/ember-1. Use that source for architecture, training, release and evaluation details.

What models should I compare Ember-1 against?

Common Ember-1 comparisons include Ember-1 vs Claude Mythos Preview, Ember-1 vs GPT-5.1 Thinking, Ember-1 vs GPT-5.1 Instant. Compare them side by side for benchmark scores, pricing, context window, latency and provider availability.