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 Beam8G
- 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 Beam8G
565
1,100
8,600
2,514,251
Beam leads in inference efficiency for its model class -- 3-4x more efficient than GLM 5.2 and over 4x more efficient than leading Western open models.
This means Beam completes tasks faster and cheaper.
This means Beam completes tasks faster and cheaper.
Oct 5, 2026 · 7:11 PM UTC
11
32
610
88,418
We scaled Beam with the largest publicly documented RL run we’re aware of, which ran on 10.5k GB300s for 4 weeks.
Algorithmic advances coupled with distributed infra enabled an RL system that scales.
Across our eval suite, capabilities continued to improve as we increased RL - u.
Algorithmic advances coupled with distributed infra enabled an RL system that scales.
Across our eval suite, capabilities continued to improve as we increased RL - u.
11
26
446
58,263
Sustained RL improvement was made possible by a strong foundation for reasoning.
Beam was pretrained in 4 weeks on 24T high-quality tokens to have innate coding capabilities.
Innovations in MoE stability and large-scale data curation & deduplication enabled us to produce a founavailable open-source base models of the same class.
Beam was pretrained in 4 weeks on 24T high-quality tokens to have innate coding capabilities.
Innovations in MoE stability and large-scale data curation & deduplication enabled us to produce a founavailable open-source base models of the same class.
5
14
332
37,322
Beam combines strong agentic performance, efficient reasoning, and a 500B form factor to give enterprises, governments, and developers a true workhorse open model. It's our first step toward building frontier intelligence that is open and widely accessible to all.
Now in the final stages of red-teaming, Beam will be released under an Apache 2.0 license this month, along with quantized FP8 and NVFP4 numerics for efficient deployment.
Sign up for early access: platform.reflection.ai
Now in the final stages of red-teaming, Beam will be released under an Apache 2.0 license this month, along with quantized FP8 and NVFP4 numerics for efficient deployment.
Sign up for early access: platform.reflection.ai
15
21
439
31,073
@reflection_ai the 4x excludes prefill and serving overhead
0
0
1
340
@reflection_ai Very impressive! Not an easy metric to optimize for
0
0
1
197
@reflection_ai Hey @reflection_ai 7Xtuw6jiur5Csu7L69KBAZKzEpr7o2sTHjTFP5bRpump
Stealth launch right ticker right name right everything?
Stealth launch right ticker right name right everything?
0
0
0
7
@reflection_ai The inference efficiency gains for Beam are really impressive
0
0
0
76
@reflection_ai Please remember to label which qwen 3.8, here I assume it’s 27B
1
0
0
273