Fireworks: Ember-1

FIREWORKS Developer Architecture Profile

OpenRouter ↗HuggingFace ↗

Intelligence (ELO)1508Chatbot Arena Verified

Max Context Limit1,048,576Max Out: 943,718 tokens

Prompt Caching90% OFF$0.30 / 1M cached

Standard API / 1M$18.00🧠 Deep Reasoner

Model Capabilities & Contract SLAs

Reasoning
Coding & Logic
Fictional
🧠 Extended Test-Time Reasoning

Ember-1 is a specialized reasoning model from Fireworks Research, built on [Kimi K3](https://openrouter.ai/moonshotai/kimi-k3). It is designed to make every token go further: it produces shorter reasoning traces, using roughly 40%...

Granular Pricing Matrix

Input Tokens (Standard Prompt)$3.00 / 1M

⚡ Cached Input (Prompt Cache Read)90% OFF$0.30 / 1M

Output Tokens (Completion)$15.00 / 1M

Pricing data via OpenRouter. Sync: 10/2/2026

Evaluate Competitors

VS Engine MatchupFireworks: Ember-1 vs Google: Gemini 3.5 Flash Lite (batch)VS Engine MatchupFireworks: Ember-1 vs OpenAI: GPT-5.5 Pro VS Engine MatchupFireworks: Ember-1 vs OpenAI: GPT-5.4 Mini (batch)VS Engine MatchupFireworks: Ember-1 vs OpenAI: GPT-5.4 Pro VS Engine MatchupFireworks: Ember-1 vs OpenAI: GPT-5.2 Pro VS Engine MatchupFireworks: Ember-1 vs OpenAI: GPT-5.2 Pro (batch)