- 2Releases tracked
- Oct 16Latest · Beam 1.0 (open weights)
- 50%Open-weight
- USUnited States
Reflection AI is a US-based AI company that builds open-weight models and an open software stack that organizations can inspect, control and deploy. Its first model, Beam, is a 501B-total, 23B-active mixture-of-experts model for coding, reasoning and agentic work, previewed in early access on October 5, 2026. Apache 2.0 weights followed on October 16 in BF16, FP8 and NVFP4, alongside a speculative-decoding drafter and a public-beta API. Reflection also offers deployment from private cloud to on-prem, air-gapped and edge environments, plus a forward-deployed engineering team.
Reflection AI
2 releases latest Oct 16, 2026
-
Beam 1.0 (open weights)
Apache 2.0 weights for the 501B-total / 23B-active MoE in BF16, FP8 and NVFP4, plus a speculative-decoding drafter; public-beta API.
-
Beam
501B-total / 23B-active MoE for coding, reasoning and agents; launched as an early-access preview. Apache 2.0 weights released Oct 16.
Reflection AI FAQ
- What is Reflection AI's latest AI model?
- Reflection AI's most recent release in the Startrise tracker is Beam 1.0 (open weights), released Oct 16, 2026.
- How many AI models has Reflection AI released?
- Startrise tracks 2 notable Reflection AI releases since January 2025, from Beam (Oct 5, 2026) to Beam 1.0 (open weights).
- Does Reflection AI release open-weight models?
- Yes. 1 of 2 tracked Reflection AI releases (50%) ship with downloadable weights; the most recent is Beam 1.0 (open weights).
- Where is Reflection AI based?
- Reflection AI is based in United States.