Reflection AI open-sources Beam: 501B-param MoE claiming 3-4x inference efficiency over GLM-5.2
rohanpaul_ai · x · 2026-10-07
Reflection AI released Beam, an open agentic MoE model with 501B total / 23B active parameters, trained end-to-end from scratch; full weights drop this month.
Key facts:
- Pretraining: 23.8T tokens in under 4 weeks on 6,144 GB300 GPUs; 1M-token effective context
- RL run: 10.5K GB300 GPUs for 4 weeks, 100M+ rollouts across 1M environments and 1.3B sandboxes, with no sign of a plateau
- Claimed strongest Western open model: 80.1 on Terminal Bench v2.1 vs GLM-5.2's 81.0, but trails DeepSeek V4.1 Flash (90.6) and Kimi K3 (88.3)
- The touted 3-4x efficiency edge over GLM-5.2 rests on estimated forward-pass FLOPs, excluding prefill, attention and serving overhead — not a measured cost
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