A newer version of this model is available: reflection/Beam-1.0-501B-A23B

Cat Jam Free, an ivory-white cat concentrating on the beat beside 1M context and free-preview labels.

Our token-count test matched Cat Jam to MISSINGNO. on all 24 phrases, a day before Reflection named both as previews of Beam.

By Stealth Models.

StealthMark results for the released Beam 1.0:

StealthMark 1.8.8 results for the released Beam 1.0: benchmark score 31.9 out of 100, General Intelligence 30.9, Visual Intelligence 33.8. Beam, the model behind Cat Jam, is strongest at composing a whole scene: 52.4 on Beach.

Every StealthMark result of the released Beam 1.0: Beach 52.4, Foresight 49.4, Raccoon 46.3, Selfie 45.4, Latch 37.3, Ground Truth 27.9, Newton's Mirror 21.7, Mona Lisa 0.7.

See the token-count comparison

The case file. First seen 13 Oct 2026, revealed 16 Oct 2026, developer Reflection AI, context 1,048,576 tokens, input Text, tool calls Yes, speed in our tests unknown. 13 Oct 2026: Appears as a stealth model for fast agentic coding, with no developer named. 15 Oct 2026, our finding: Our token-count test matches its 24 counts to MISSINGNO. 16 Oct 2026: Reflection names the Cat Jam demo as a Beam preview and releases Beam 1.0 under Apache 2.0.

Compare Beam with MiMo-V2.6-Flash

Facts as text
Fact Value
First seen 13 Oct 2026
Revealed 16 Oct 2026
Developer Reflection AI
Context 1,048,576 tokens
Input Text
Tool calls Yes
Speed in our tests unknown
Access Preview ended; the released Beam is available

Run it

Beam 1.0 is open under Apache 2.0: reflection/Beam-1.0-501B-A23B. Even in 2-bit it needs about 130 GB. GGUF files: unsloth/Beam-1.0-501B-A23B-GGUF.


Model guide · StealthMark leaderboard · Follow us
Text and images: CC BY 4.0, stealthmodels.com.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Space using stealth-model/cat-jam 1