Reflection unveils Beam: 501B-param open MoE model with 23B active, trained on 23.8T tokens
bigblueboo · x · 2026-10-06
Reflection AI announced Beam, its first open-weight model, with full weights, technical report and model card due later this month.
- Architecture & training: sparse MoE with 501B total / 23B active parameters; pretrained from scratch on 23.8 trillion curated tokens
- Heavy RL investment: a 4-week high-compute RL run on 10.5K NVIDIA GB300 GPUs generated over 100 million rollouts
- Positioning: focused on coding, reasoning and agentic workloads; claims to advance the Western open-weight frontier, competitive with GLM 5.2 and approaching Qwen 3.8-Max on coding/agentic tasks, with inference efficiency as its edge over stronger open models like Kimi K3
- Currently in final red-teaming; early access sign-ups are open
More from Models
- Jeremy Howard: Google shipped the best model ever made; OpenAI its worst, same day — hugobowne · 2026-11-20
- Polymarket puts 83% odds on Google having the best AI model by year-end — Polymarket · 2026-11-20
- One year later, model naming is still a mess: the gpt-5.1-codex-max anniversary — heyneighbor · 2026-11-20
- Developer slams Grok models as 'garbage', shares coding agent's repetitive trial-and-error log — feregri_no · 2026-11-20
- Drex Router cuts small-game build cost to $1.26 vs $13.44 on Claude Opus alone — testingcatalog · 2026-11-20
- OpenAI slammed for AI slop: 'Just hire photographers to RLHF imagegen for a day' — flowersslop · 2026-11-20