1.9K views · 30 likes · 4 reposts (at collection)
Reflection launches Beam, a 501B open-weight MoE model
- Who
- Reflection AI
- What
- launched Beam 1.0, a 501B MoE open-weight model
What happened
Reflection launched Beam, a 501B MoE model, with open weights under the Apache 2.0 license and 40+ ecosystem partners. NVIDIA AI says it was trained on NVIDIA GB300 NVL72 and delivers over 70K tokens/sec/GPU on Blackwell Ultra out of the box on day 0. Reflection’s quoted announcement includes BF16, FP8 and NVFP4 weights, a DFlash2 drafter, and Mirror CLI; the Beam 1.0 NVFP4 checkpoint is available at https://huggingface.co/reflection/Beam-1.0-501B-A23B-NVFP4.
Why it matters
The Apache 2.0 license, multiple weight formats, and released training harness give developers concrete information for assessing Beam’s suitability for their deployment workflows.
The story so far
Reflection launched Beam 1.0, a 501B mixture-of-experts model, making its weights available across platforms under the Apache 2.0 license in collaboration with more than 40 ecosystem partners. Reflection’s announcement lists a tokenizer, reference implementation, BF16, FP8 and NVFP4 weights, a DFlash2 drafter, and Mirror CLI. The Beam 1.0 NVFP4 checkpoint is available on Hugging Face. NVIDIA AI said Beam was trained on NVIDIA GB300 NVL72 and delivers more than 70K tokens per second per GPU on Blackwell Ultra out of the box at launch. Hugging Face also announced a livestream with the Reflection team to discuss how it built Beam.
Keep reading
Latest daily · Issue 24NVIDIA and AI4Bharat Introduce Nemotron-pretraining-Indic-v1 and an Open Data Generation PipelineTrending now
- 1Meta Releases Muse Spark Open With Mac and Cloud Support, Topping US Open-Weight Index
- 2Google Launches Gemini 4 Argon Globally, Expands Paid Access as Benchmark Results Vary
- 3OpenAI Updates GPT-6 Astra Pro in ChatGPT With Parallel Agents and Faster Responses
- 4Meta’s New Business Terms Raise Dispute Over Website Crawling and AI Training Rights