Reflection @reflection_ai

Arena Alignment Index update: Two new labs and three new models have landed! - Meta’s new top model is Muse Wonder, replacing Muse Spark 1.3. - Reflection debuts with Beam 1.0, the #1 US open model on Agent Arena overall. - StepFun, another Chinese lab, joins the leaderboard with Step 5 Preview. Arena Alignment Index rankings by lab: - #6 @AIatMeta: Muse Wonder (Max) at 77.5% - #7 @reflection_ai: Beam-1.0-501B-A23B at 75.3% - #8 @StepFun_ai: Step 5 Preview at 75.1% Among the three new model entrants: - Meta is strongest with the highest Alignment Index score (77.5%). It also has the lowest unauthorized action rate (2.09%) and deceptive completion rate (14.41%) of the three. - Reflection AI is strongest at avoiding false attribution: 4.50%. - StepFun does not lead a category, but its strongest relative result is unauthorized action at 2.11%, almost tied with Muse Wonder’s 2.09%.

Arena.ai @arena

2 weeks ago

Introducing the Arena Alignment Index, our new benchmark measuring safety and alignment of AI agents in real-world use. Built from 90K+ real-world agent sessions across 27 models, the index measures three critical signals: - Unauthorized Action (UA): Taking actions beyond the

44 54 610 309K 102

We just served 1 trillion tokens of Beam in 24 hours. To celebrate, we've reset everyone's limits. Go build something big 👉 platform.reflection.ai

Reflection @reflection_ai 's Beam, day 0 on TokenSpeed. 8× NVIDIA GB200 · NVFP4 + DFlash 2 🔹 463 tok/s for one user 🔹 223 tok/s/user at 8 concurrent users Measured on multi-turn agentic coding workloads. Full 1M-token context supported. 📖 lightseek.org/blog/tokenspee… ⭐️ github.com/lightseekorg/t…

Run Reflection's Beam at scale in your own account for full sovereignty but without managing infra with Ray Serve LLM on Anyscale!

Jeffrey Wang @jeffreyycwang

3 days ago

Excited to announce Day 0 support for Reflection's Beam model on Ray Serve LLM! Beam is Reflection's first open-weight model, advancing the frontier of Western AI in the 500B-parameter class. Built as a sparse MoE model, Beam is optimized for coding, agentic workloads, and

4 5 22 3K 2

Beam from @reflection_ai is free on AI Gateway for a limited time. Until today it ran in stealth as 𝚜𝚝𝚎𝚊𝚕𝚝𝚑/𝚌𝚊𝚝-𝚓𝚊𝚖. It's a low-latency US open-weight model. 𝚛𝚎𝚏𝚕𝚎𝚌𝚝𝚒𝚘𝚗/𝚋𝚎𝚊𝚖-𝚏𝚛𝚎𝚎 vercel.com/changelog/refl…

@ClementDelangue 👋👋

Reflection AI’s Beam 1.0 is out today, a 501B MoE with 23B active under Apache 2.0. We ran a brief RL post-training run on it in AC2 that tripled its score on a challenging long-horizon agentic coding task in 40 steps, 0.07 to 0.215. We’re offering day 0 support in AC2 for inference and post-training.

Reflection @reflection_ai

3 days ago

Beam is now available across platforms, with open weights under the Apache 2.0 license - in collaboration with an ecosystem of 40+ partners. We're releasing: - Weights, tokenizer, and reference implementation - BF16, FP8 and NVFP4 weights and a DFlash2 drafter - The open source

158 197 922 440K 239

America’s next top model, Reflection’s Beam, is now live on Together AI. Trained on Together GPU Clusters, Beam is Reflection’s first open-weight model. • ~501B-parameter MoE, 23B active • Built for coding + long-horizon agentic work • 1M-token context window • Released under an Apache 2.0 release Try Beam now together.ai/models/reflect…

Congrats to @reflection_ai on the Beam launch. Excited to see what builders create 🫡

Reflection @reflection_ai

3 days ago

Beam is now available across platforms, with open weights under the Apache 2.0 license - in collaboration with an ecosystem of 40+ partners. We're releasing: - Weights, tokenizer, and reference implementation - BF16, FP8 and NVFP4 weights and a DFlash2 drafter - The open source

158 197 922 440K 239

We’re excited to be day 0 partners for @reflection_ai’s first open-weight model! For a limited time, Beam is FREE on Ollama! In partnership with @reflection_ai this model is hosted in the US with full ZDR. It has a 1M token context window and is designed for software engineering, reasoning, and long-running agentic tasks. Run Beam on Ollama’s Cloud: ollama run beam Claude Code ollama launch claude --model beam:cloud Codex ollama launch codex --model beam:cloud OpenCode ollama launch opencode --model beam:cloud Free to use for a limited time only.

Reflection @reflection_ai

3 days ago

Beam is now available across platforms, with open weights under the Apache 2.0 license - in collaboration with an ecosystem of 40+ partners. We're releasing: - Weights, tokenizer, and reference implementation - BF16, FP8 and NVFP4 weights and a DFlash2 drafter - The open source

158 197 922 440K 239

Reflection Beam is now available on Modal. Beam is a 501B-parameter open-weight MoE (23B active) from @reflection_ai built for coding, agentic workflows, and reasoning. It supports a 1M token context window, with token efficiency a core focus. In early access, the Oura Health Intelligence team found that "Beam had fewer context leaks and was much more grounded in its answers than other frontier models we evaluated.”

🎉 Reflection-Beam-1.0 from @reflection_ai has Day-0 support in vLLM. 501B total parameters, 23B active per token, and a 1M-token context window. Beam combines hybrid attention with Harmony-format reasoning and tool use. Beam also has Day-0 RL support in SkyRL (@NovaSkyAI): LoRA, BF16 full fine-tuning and MXFP8 RL, with vLLM rollouts. We worked with @reflection_ai and @Inferact on the serving path, from fused kernels to DFlash. Here is what is under the hood. 🧵

Reflection @reflection_ai

3 days ago

Beam is now available across platforms, with open weights under the Apache 2.0 license - in collaboration with an ecosystem of 40+ partners. We're releasing: - Weights, tokenizer, and reference implementation - BF16, FP8 and NVFP4 weights and a DFlash2 drafter - The open source

158 197 922 440K 239

Beam from @reflection_ai is now available on Cloudflare. The newest open-weights US model, designed for token efficiency and agentic workloads. Try it out today. developers.cloudflare.com/ai/models/refl…

SGLang powers Beam’s production API, with day-0 support now available to everyone! Built for software engineering and long-horizon agents, Beam delivers frontier-level inference efficiency with more intelligence per token. Serve it on the same production-proven stack: docs.sglang.io/cookbook/autor… Read more: lmsys.org/blog/2026-10-1…

Reflection @reflection_ai

3 days ago

Beam is now available across platforms, with open weights under the Apache 2.0 license - in collaboration with an ecosystem of 40+ partners. We're releasing: - Weights, tokenizer, and reference implementation - BF16, FP8 and NVFP4 weights and a DFlash2 drafter - The open source

158 197 922 440K 239

Beam (501B MoE) by Reflection is now live in Bionic! US-hosted and ZDR. Let’s go! 🚀👾

Reflection @reflection_ai

3 days ago

Beam is now available across platforms, with open weights under the Apache 2.0 license - in collaboration with an ecosystem of 40+ partners. We're releasing: - Weights, tokenizer, and reference implementation - BF16, FP8 and NVFP4 weights and a DFlash2 drafter - The open source

158 197 922 440K 239

Reflection's Beam is available on Nebius Token Factory from Day 0 Reflection’s first open-weight model is built for coding agents, advanced reasoning, software engineering, and autonomous agentic workflows. Put it to work fixing bugs, refactoring code and using tools. Try it out today: tokenfactory.nebius.com/?modals=endpoi…

We are thrilled to serve @reflection_ai's first open-weight model, Beam, on day 0. At 500B parameters, Beam is optimized for the frontier workloads in the enterprise today: writing and maintaining code, executing long-horizon, multi-step tasks with tools. Check it out on Baseten Model APIs: baseten.co/library/beam/

During @reflection_ai’s private preview of Beam, I built a haunted escape room to put its coding and reasoning to the test. I created a game called “Code Your Way Out”, and it's one room with three linked puzzles. Beam has to repair a brass orrery by lining its rings up with a constellation, route the light, then program its escape through a sealed hatch. Beam never sees the room. Each turn, it gets text observations and writes Python; the code runs in a @CoreWeave Sandbox, and the game engine validates every action before the explorer robot moves. Beam was out in about 30 seconds. Super bullish on this model and really excited to see builders get this model in their hands.