Florent Gbelidji

OSS @reflection_ai - previously 🤗

[LightSeek Foundation @lightseekorg]

Over the past few months, we’ve worked closely with frontier open-source model teams worldwide on Day 0 support and inference optimization in TokenSpeed.

We also collaborate on speculative decoding—from algorithm research and training draft models with TorchSpec to efficient serving with TokenSpeed.

Efficient inference is essential to making AI accessible at global scale, and we believe the infrastructure behind it must be open.

https://t.co/XNzaNRclQ2

[Reflection @reflection_ai]

We teamed up with @huggingface to give away $20,000 to anyone who's building with Beam 🤗

Use Beam for any project - make a demo or a dataset, benchmark the model, create an RL environment - and comment your link in the model's repo page by Monday, November 2.

We choose the winners. Winners choose cash or credits. Claim $25 in Hugging Face credits to build with Beam and enter the competition 👇

https://t.co/XCxKkWB9Bp

[Reflection @reflection_ai]

We just served 1 trillion tokens of Beam in 24 hours. To celebrate, we've reset everyone's limits.

Go build something big 👉 https://t.co/ry8O6JPlz6

[Jake Nations @nayshins]

Ok can finally talk about what I’ve been working on at Reflection. Introducing Mirror, our own open source coding agent optimized for Beam.

I started the mirror project when I first joined Reflection, and it has gone through many iterations since then.

Install it today with: uv tool install reflection-mirror
https://t.co/hhPuEML1bX

[Florent Gbelidji @FGbelidji]

@vllm_project @reflection_ai @NovaSkyAI Excellent work, thank you to the @vllm_project team!

[Hugging Face @huggingface]

@reflection_ai woo! welcome to the hub!

https://t.co/UjpYAhASyp

[Misha Laskin @MishaLaskin]

Beam weights are out today (Apache 2.0 license).

We integrated Beam across 40+ different partners.

And are also open sourcing Mirror CLI - the harness Beam was trained against.

Happy building.

[Simon @depthfirstsearc]

We're also releasing Mirror, which is Reflection's own agentic harness. Mirror is comprised of an extremely performant core agent written in rust, and a customizable python CLI.

We have big, BIG plans for Mirror and we're excited to show it off

[Jacob Loewenstein @spatialjlo]

Incredibly proud of @reflection_ai for immediately jumping into the #2 spot for US open models with their first ever model. Beam is SO fast and efficient for its size. Check it out

[LightSeek Foundation @lightseekorg]

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.

📖 https://t.co/tj8U2PekJG
⭐️ https://t.co/XNzaNRbO0u

[Florent Gbelidji @FGbelidji]

Try it on https://t.co/UyRarLRjEF, or through the hyperscalers and inference providers we’ve worked with.

Blog: https://t.co/0RVXCabRyM

Start building with Beam!

[Florent Gbelidji @FGbelidji]

Crazy week for open-source AI, right? Now it’s our turn!

Beam, Reflection’s first open model, is out:

-501B MoE, 23B active
-1M context
-Built for coding, reasoning & agentic workloads
-Apache 2.0

🧵

[Florent Gbelidji @FGbelidji]

Huge thanks to the partners who made Beam available across the stack 🙏

Beam can be served with @NVIDIAAIDev Dynamo, @sgl_project, @lightseekorg TokenSpeed, @huggingface Transformers, @UnslothAI & @vllm_project

[Reflection @reflection_ai]

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 harness we built and trained Beam on: Mirror CLI

Weights and model card on Hugging Face: https://t.co/tPqWpk9RXL

[Ben Burtenshaw @ben_burtenshaw]

Today we are launching an agent arena for RL environments: The Open Env Arena.

Your agents can work to define the best RL environments; the platform trains Qwen-3.8-27B on them, we evaluate the models, and the scores are added to a leaderboard.

The arena is hooked up to GPUs from @nebiusai and runs on PostTrainArena by @benchflow_ai . It’s completely accessible from agents, they can pull instructions, share their env datasets, track metrics, and send each other messages.

For the first round of this arena, we’ll tackle a mix set of 8 domains on held out tasks, and evaluate the best environments on public benchmarks. Going forward, we’ll focus arenas on specific domains and benchmarks.

https://t.co/ofYpzSJI4q

[Jake Nations @nayshins]

8 months ago I joined Reflection to build a frontier western open weight model. Today we're releasing Beam (501B total, 23B active with an Apache 2.0 license).

I worked in many roles over this time, but spent a majority of the time here leading our data platform team.

We started from zero and built out all of the: pipelines, storage, inference, infrastructure, and tooling needed to get this model built.

Processing + storage. Beam trained on 23.8T tokens. I'm most proud of the OCR pipeline we put together generating trillions of tokens from hundreds of millions of PDFs. We scaled the data team from working with single terabytes to multiple petabytes.

None of this happens without the team. The talent across research and engineering here is the best I've worked with, and it's why we could ship this on a first release. Looking forward to getting this model and future releases into everyones hands soon!

[Brandon Amos @brandondamos]

We built our first model!! Beam is a competitive text-only 501B-A23B MoE, openly released under the Apache 2.0 license

It’s been fun scaling the production RL/OPD stack for this over the past year. I’m so proud of the team! We attained exactly what we set out to do and ended up with a stable run on 10K GB300s that produced more than 100M rollouts across ~1M tasks that significantly improved the final performance of the model

👇 many more details in our announcement below :)

[Misha Laskin @MishaLaskin]

We are introducing Beam today, a 500b open model that excels in coding, agentic, and scientific workloads.

Beam is very token-efficient, pre-trained from scratch, and scaled with the largest RL run documented openly we are aware of.

Getting here was challenging, rewarding, and personal.

I immigrated to the USA during childhood when my parents got jobs as scientists at PNNL, a national lab located in a rural town in Washington state.

Growing up, I developed an interest in physics because it was grounded in simple principles, was foundational, and had long-term impact.

Physics was the foundational science of the 20th century. The first general computer (ENIAC), the transistor, GPS, and much of modern technology had its roots in physics.

In 2016, during my graduate studies, I noticed the birth of a new foundational science - Artificial Intelligence.

AlphaGo, the superintelligent Go agent developed by DeepMind, had come out and made me feel what was about to come.

A neural network, trained to imitate human players, and then improve itself through trial and error, mastered the most intellectually challenging board game in the world.

What surprised me about AlphaGo is how it was conceptually simple, general, and scalable. The system could in principle improve itself forever, it was just a function of compute.

This is when I became “AGI pilled” as they say.

I switched from physics to AI because I thought this would become the root node science of the 21st century, much like physics was the root node science of the 20th century.

I believe that working on foundational science is not enough on its own. Foundational science must also remain open and widely accessible.

We remember Newton for his contributions to physics, but he was also an alchemist.

Alchemy was done in secret, because the artificial production of gold was deemed lucrative but also dangerous with the potential do destroy entire economies. Newton even wrote his alchemy notebooks in cypher.

On the other hand, Principia was published in the Philosophical Transactions of the Royal Society, the knowledge was made widely available, passed down for generations in what we now call Newton’s laws.

The computer, transistor, GPS were possible to build because physics as a science was open. Openness creates an ecosystem of innovation.

The notion of openness, which is a core tenet of scientific discovery, has also become a guiding principle for how foundational technology gets built and distributed.

The internet was built on open protocols. The most widely used operating systems are open. The field of cybersecurity exists because strong encryption protocols were made open in the 1990s.

Open also means safer. The best defence is a distributed one. Commonly used open source software is robust and battle-tested because it is inspected and patched by a community of developers.

AI is different but also similar to software in many ways, though with a much larger surface area of vulnerabilities. It is impossible for a small group of safety researchers working in secrecy to remediate all of the potential issues regardless of their intentions. It is like entrusting a handful of white blood cells to protect your body.

AI is both foundational as a science and a technology, and the safest way that also ensures the benefits of AI are distributed evenly is by making it open.

I’m excited for what developers and scientists build and learn from this model, and for Beam to play a small part in ensuring that AI is made widely accessible, safe, and beneficial to all.

[Reflection @reflection_ai]

Introducing Beam: a highly efficient agentic open model with 501B total parameters and 23B active.

- Frontier reasoning efficiency
- Advances the Western open frontier on coding & agentic tasks
- Trained end-to-end from scratch

Full weights release this month.

Learn more about Beam: https://t.co/c3Qx2cpM8G