Fournisseur

Reflection

Reflection AI is best evaluated through Beam, a large open-weight model built for enterprise coding and agentic workloads.

summary

Reflection AI is an American AI company based in New York, founded in March 2024 by former Google DeepMind researchers Misha Laskin and Ioannis Antonoglou. In October 2026 it released Beam, its first open-weight model: a mixture-of-experts model with 501 billion total parameters and 23 billion active per token, aimed at enterprise coding and agentic workloads. On Eden AI, Beam is available with reasoning, function calling and a context window of about one million tokens, served by several providers.

What is Reflection AI?

Reflection AI is an American artificial intelligence company based in New York. It was founded in March 2024 by Misha Laskin and Ioannis Antonoglou, two former researchers at Google DeepMind. The company focuses on open-weight frontier models that businesses and governments can run and control themselves.

Reflection models available on Eden AI

Beam: Reflection's first open-weight model, released in October 2026. It is a mixture-of-experts model with 501 billion total parameters and 23 billion active per token, with reasoning, function calling and a context window of about one million tokens.

When to choose Reflection AI

You need a large open-weight model for coding and agents.
You want a Western alternative to large open models from Chinese labs.
You want very long context for big codebases or documents.

How to use Reflection AI with Eden AI

Call Beam with the Eden AI chat completions API and your Eden AI API key. The model page lists the providers serving it, with prices, regions and measured speed.