A 501B-parameter open-weight model from a US company, comparable to Chinese open models and slightly behind on agentic coding. Only available to early testers.
Reflection Beam
Reflection AI's 501B-parameter open-weight mixture-of-experts model for coding, reasoning and agentic workloads.
Updated EU status checked Last shown
Reflection Beam at a glance
- What is Reflection Beam?
- Reflection Beam is a model from Reflection AI. Reflection AI's 501B-parameter open-weight mixture-of-experts model for coding, reasoning and agentic workloads.
- Who makes Reflection Beam?
- Reflection Beam is made by Reflection AI.
- Does Reflection Beam work in the EU?
- Available in the EU. Open-weight model deployable on-prem, in private cloud and on sovereign infrastructure via partners, and Reflection's privacy policy has a European Region version for EEA users. No explicit EU exclusion was found, though no formal supported-countries list was seen. Last verified 14 Nov 2026.
- How much does Reflection Beam cost?
- API pricing. Served through Reflection's API platform and partner managed endpoints; pricing not stated in the provided material.
- Where was Reflection Beam shown?
- In 1 creator video by Matt Wolfe. Most recently by Matt Wolfe on 9 Oct 2026, at 25:26 in “AI News: Haiku-5.5, Visual ChatGPT, Grok's New Trick, and more!”.
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
Beam is the first open-weight model from Reflection AI, a US company focused on open models. It is a 501B-parameter mixture-of-experts (MoE) model built for coding, reasoning and agentic workloads. Creator coverage describes it as comparable to Chinese open-weight models and slightly behind them on agentic coding.
Reflection says it works with partners across the open ecosystem on inference, quantization, fine-tuning, reinforcement learning, agent harnesses, evaluation and observability. Partners make Beam available through managed endpoints, cloud platforms, on-premises systems and sovereign infrastructure. Reflection also offers its own API platform and deployment into private cloud, on-prem, air-gapped and edge environments, plus forward-deployed engineering support.
The provided material does not state the license, exact release date, pricing or context length. One early creator report said access was limited to early testers, while the vendor's blog title ("Beam is available") suggests broader availability later.
- 501B-parameter mixture-of-experts architecture
- Open weights that can be inspected, fine-tuned and deployed
- Aimed at coding, reasoning and agentic workloads
- Available via Reflection's API, partner endpoints, cloud, on-prem and sovereign infrastructure
- Supported by partners for inference, quantization, fine-tuning and RL
As seen on YouTube
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Similar AI tools
Closest neighbours by meaning, not just by category.
Agents A1
Model
35B Mixture-of-Experts open-weights model from InternScience for long-horizon agentic search, engineering and tool use.
Muse Spark Open
Model · Meta
Meta's open-weight 816B-parameter mixture-of-experts model (42B active, 1M context) in the Muse model family.
LongCat 2.0
Model · Meituan
Meituan's open-weights 1.6T-parameter mixture-of-experts LLM (48B active), aimed at agentic coding and reasoning tasks.
Occamy 1.0
Model · Accio Lab
Open-weights 35B MoE (3B active) agentic model from Accio Lab, post-trained from Qwen3.6-35B-A3B.
Sage A1.5
Model · Alpha Innovator
35B mixture-of-experts open-weights agentic model for engineering, long-horizon search and scientific research.
Ripple 65B Preview
Model · Recursive Intelligence
Dense 65B open-weight language model from Ricursive Intelligence, trained for agentic coding with up to 1M context.