# Reflection AI — Beam 1.0

> AI lab announcements, news reports and podcasts.

forck.live follows what companies announce about their own products. Press may
corroborate an official story; it never leads and never appears as an item of its
own. A missing date stays missing. A title is the publisher's words — where a
company does not publish in English, an English rendering is listed separately
as `Title in English` and named as this site's translation, never as theirs.

This file is the record of one model version: what the model registry states
about it, Reflection AI's launch post, the first-party posts this record
holds about it, the outlets that reported on those posts, and other companies'
announcements that it runs on their platforms. A model ranks nothing here and
nothing is composed: each date names whose it is, and a field nothing states is
`not stated` rather than missing.

- This model's page: https://forck.live/models/beam-1-0-501b-a23b
- Reflection AI's page: https://forck.live/company/reflection-ai
- Every model: https://forck.live/models
- Catalogue: https://forck.live/llms.txt

## Model

- Name: Beam 1.0
- Company: Reflection AI (reflection.ai)
- Launch post: Introducing Beam: Reflection’s 501B open-weight model — https://forck.live/items/18513-introducing-beam-reflection-s-501b-open-weight-model
- Announced: 2026-10-05
- Launch source: https://reflection.ai/blog/introducing-beam
- Launch chosen by: a person, pinned in the model registry
- Released (per models.dev): not stated
- Listed (per OpenRouter): 2026-10-16
- Open weights: not stated
- Posts about it: 1
- Outlets named under its posts: 25
- Also written: Beam-1.0
- OpenRouter id: reflection/beam-1.0-501b-a23b — https://openrouter.ai/reflection/beam-1.0-501b-a23b
- Context window: 256K tokens — from https://forck.live/items/18513-introducing-beam-reflection-s-501b-open-weight-model

## Posts about it

_Reflection AI's own posts whose title names Beam 1.0, and its launch post, in the record's order: dated newest first, undated last. The record reads the newest 50 first-party posts mentioning the model, every company's together, so a model mentioned more often than that lists the newest of them._

Every entry is one first-party announcement, and carries the same eight fields
in this order before any others:

    Company · Announced · Category · Coverage · Announcement · Group · Source · Record

`Source` is the company's own page the announcement was traced to and `Record`
is this site's page for it. Both are safe to cite.

`Coverage` is a number of outlets, or `not counted` — which means this record
reached no count of its own, never that nobody wrote about it.

`Category` is this tracker's own analysis, and is authoritative wherever it is
stated. `not stated` means either that nothing has analysed the post yet or
that the analysis placed it outside this taxonomy; neither is a verdict about
the announcement.

`Announcement` is whether the post is news of something the company did, rather
than a how-to, a customer story, a changelog listing or a staffing note.

`Group` is which of this record's groups the entry falls in, carried on the
entry itself so it holds where the groups are headings and where they are not.
One of: `models`, `covered`, `announcements`, `routine`.

Anything after `Record` is held detail that only some announcements carry.

### Reflection AI — Introducing Beam: Reflection’s 501B open-weight model

- Company: Reflection AI (reflection.ai)
- Announced: 2026-10-05
- Category: open-weight-release
- Coverage: not counted
- Announcement: yes
- Group: models
- Source: https://reflection.ai/blog/introducing-beam
- Record: https://forck.live/items/18513-introducing-beam-reflection-s-501b-open-weight-model
- Subject: Reflection Beam

Reflection AI introduced Beam, an open-weight sparse Mixture-of-Experts model with 501 billion total parameters and 23 billion active parameters, focused on coding, reasoning, and agentic tasks. The model was pretrained on 23.8 trillion tokens and trained with high-compute reinforcement learning using 10.5K NVIDIA GB300 GPUs over four weeks, generating over 100 million rollouts. Beam is undergoing final red-teaming and evaluations, with weights, technical report, and developer artifacts to be released later this month.

## Who reported on it

_Outlets this record can name and link, under the first-party post each one covered. Their number is `Outlets named under its posts` above, and the rows are not totalled again. A post's own `Coverage` can be higher: it counts outlets the signal saw that this record cannot name and link._

### On: Introducing Beam: Reflection’s 501B open-weight model

Record: https://forck.live/items/18513-introducing-beam-reflection-s-501b-open-weight-model

- The Decoder — Reflection's Beam becomes the most capable open-weight model built outside China (2026-10-06T12:11:22+00:00): https://the-decoder.com/reflections-beam-becomes-the-most-capable-open-weight-model-built-outside-china/
- SiliconANGLE — Reflection AI debuts open-source Beam model with 501B parameters (2026-10-06T00:46:33+00:00): https://siliconangle.com/2026/10/05/reflection-ai-debuts-open-source-beam-model-with-501b-parameters/
- TechCrunch — Reflection debuts Beam, an open-weight AI model to rival Chinese models at lower compute cost (2026-10-05T19:33:53+00:00): https://techcrunch.com/2026/10/05/reflection-debuts-beam-a-open-weight-ai-model-to-rival-chinese-models-at-lower-compute-cost/
- Axios — "I don't want an oligopoly": New open-weight AI models mount a comeback against China: https://axios.com/2026/10/06/reflection-mistral-open-weight-ai-models-china
- Bloomberg: https://bloomberg.com/news/articles/2026-10-05/nvidia-backed-reflection-unveils-open-ai-model-taking-on-china
- Business Standard: https://business-standard.com/technology/artificial-intelligence/nvidia-backed-reflection-unveils-open-ai-model-to-rival-chinese-models-126100600175_1.html
- Cautious Optimism — Splitting the atom to juice Western open-weight models: https://cautiousoptimism.news/atom-split-western-ai
- Finimize: https://finimize.com/content/nvidia-backed-reflection-ai-unveils-beam-to-take-on-chinas-coders
- Fortune — Reflection AI unveils its first open model, Beam. Could it be the America's best chance to compete with China?: https://fortune.com/2026/10/05/reflection-ai-unveils-beam-a-new-us-based-open-source-model-to-compete-with-china
- Futurum — Reflection's Beam: Open-Weight Frontier at Enterprise Inference Cost: https://futurumgroup.com/insights/reflections-beam-open-weight-frontier-at-enterprise-inference-cost
- Help Net Security — Reflection's Beam trails top open models on coding tests but claims lower inference compute: https://helpnetsecurity.com/2026/10/06/reflections-beam-trails-top-open-models-on-coding-tests-but-claims-lower-inference-compute
- Indiatimes — Nvidia-backed Reflection unveils first AI model to take on Chinese open models - The Economic Times: https://economictimes.indiatimes.com/tech/artificial-intelligence/nvidia-backed-reflection-unveils-first-ai-model-to-take-on-chinese-open-models/articleshow/134721941.cms
- PCMag: https://pcmag.com/news/nvidia-backed-startup-builds-open-weight-model-to-challenge-chinese-labs
- Quartz: https://qz.com/reflection-ai-beam-open-weight-model-launch-100626
- Reuters: https://reuters.com/technology/nvidia-backed-reflection-unveils-first-ai-model-take-chinese-open-models-2026-10-05
- RuntimeWire — Reflection publishes Beam benchmark scores ahead of its weight release: https://runtimewire.com/article/reflection-ai-beam-open-weight-model
- RuntimeWire — Reflection AI unveils Beam as a Western open-weight rival to Chinese models: https://runtimewire.com/article/reflection-ai-beam-open-weight-model-2
- Sources — Reflection’s founders on building a DeepSeek of the West: https://sources.news/p/reflection-founders-open-weight-beam-release
- South China Morning Post — Nvidia-backed Reflection AI challenges Chinese dominance in open-weight models: https://scmp.com/tech/tech-trends/article/3369885/nvidia-backed-reflection-ai-challenges-chinese-dominance-open-weight-models
- Tech in Asia: https://techinasia.com/news/nvidiabacked-ai-startup-reflection-unveils-beam-model
- The Hill: https://thehill.com/policy/technology/6130496-reflection-ai-releases-beam-model
- The Information: https://theinformation.com/briefings/reflection-ai-announces-first-open-source-model-beam
- The Neuron — Reflection AI’s Beam Takes On Chinese Open-Source Models: https://theneuron.ai/news/reflection-ai-beam-chinese-open-source-models
- The Straits Times — Nvidia-backed start-up Reflection unveils open AI model to rival China offerings: https://straitstimes.com/business/nvidia-backed-start-up-reflection-unveils-open-ai-model-to-rival-china-offerings
- Unite — Reflection AI Unveils Beam, a 501B-Parameter Open-Weight Model: https://unite.ai/reflection-ai-unveils-beam-a-501b-parameter-open-weight-model
- WOWTALE — Reflection AI Launches Open-Weight Model 'Beam,' Pushing the Western Open Frontier: https://en.wowtale.net/2026/10/06/235392

## Where it runs

_Other companies' own announcements naming Beam 1.0 and Reflection AI — each is that company's post, not Reflection AI's._

### Amazon — Introducing Reflection Beam: an open-weight model for agentic coding on AWS

- Company: Amazon (amazon.com)
- Announced: 2026-10-16T15:23:10+00:00
- Category: developer-tool-release
- Coverage: not counted
- Announcement: yes
- Group: announcements
- Source: https://aws.amazon.com/blogs/machine-learning/introducing-reflection-beam-an-open-weight-model-for-agentic-coding-on-aws/
- Record: https://forck.live/items/23589-introducing-reflection-beam-an-open-weight-model-for-agentic-coding-on-aws
- Subject: Bedrock / Nova

Reflection AI introduced Beam, an open-weight sparse mixture-of-experts model for agentic coding and long-horizon tasks, with 501 billion total parameters and 23 billion active parameters. The model is available in Amazon SageMaker JumpStart in two US East regions, and its weights are released on Hugging Face under the Apache 2.0 license in BF16, FP8, and NVFP4 checkpoint variants. The post provides deployment guidance for SageMaker JumpStart real-time endpoints and SageMaker HyperPod Inference.

---

Record: https://forck.live/models/beam-1-0-501b-a23b
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md