THE HOOK
A $25 billion startup finally has a model to show for it, and you'll be able to download it for free, eventually.
THE GIST
Reflection AI unveiled Beam, a text-only mixture-of-experts model built for reasoning, coding and agent work. It claims reasoning scores on par with Z.ai's GLM-5.2 while using 3-4x less inference compute. Weights land under Apache 2.0 later this month.
THE TAKE
Benchmarks you mark yourself are homework, not proof, and nobody outside Reflection has checked these yet. Still, a credible US open model is overdue, and cheaper than China is the only pitch moving enterprise buyers right now. Text-only in late 2026 is a choice, though. Ship the weights, then gloat.
OPINION
Reflection has raised roughly $4.7 billion and signed more than $7 billion of GB300 compute deals with SpaceX and Nebius, per TechCrunch, so Beam is less a launch than a down payment on those bills. The Next Web says US and UK safety bodies are still assessing it, which makes "later this month" the most important phrase in the announcement.
- Story type: Breaking news
- Impact: Medium
- Confidence: High (official announcement)
- Audience: Developers
- What changed: Reflection AI unveiled Beam, its first open-weight frontier model, with weights due this month
- What to watch: Whether the weights ship on time and independent benchmarks back the 3-4x compute claim
- Reading time: 1 minute
