PrismML

@PrismML · United States · joined 28 Mar 2025

Centering AI research on efficiency. https://t.co/88MQHGCeFD

27 806Followers

31Following

337Posts total

2.7MViews on collected posts

Son gönderiler

PrismML@PrismML17 Sep

@PrismML 9x smaller with 98.2% retention is nuts 19.1K views · 458 likes · 0 reposts · 4 replies Open on X →

PrismML@PrismML17 Sep

Building with Bonsai? We work with teams on domain-specific post-training and hardware optimization to meet memory, latency, and power constraints. Get in touch: contact@prismml.com. 74K views · 211 likes · 11 reposts · 6 replies Open on X →

PrismML@PrismML17 Sep

Computer use is another strong test. The model has to repeatedly interpret state, choose the next action, and stay coherent across a long sequence of steps, exactly where small capability gaps become visible. Here is Bonsai 2 27B running a computer-use workflow locally on the h

2:30

111.6K views · 520 likes · 23 reposts · 10 replies Open on X →

PrismML@PrismML17 Sep

Links: Blog: https://t.co/0Xh3vu12Ex Whitepaper: https://t.co/foPUfHCOV3 Models: https://t.co/M5Ymxjt9Wz WebGPU Demo: https://t.co/Xq2QAoL0KT GitHub: https://t.co/3BAp6Rg6az Docs: https://t.co/MzBfoKhxh3 Discord: https://t.co/1MwtJroLJs Join us: https://t.co/7TvnvrzO64 73.3K views · 509 likes · 38 reposts · 17 replies Open on X →

PrismML@PrismML17 Sep

On an intelligence-density basis, Bonsai 2 27B is a clear outlier relative to both full-precision models and other low-bit alternatives. Many low-bit models become deployable only with a meaningful quality tradeoff. Bonsai 2 pushes the frontier toward significantly higher https:

183.4K views · 553 likes · 21 reposts · 12 replies Open on X →

PrismML@PrismML17 Sep

This is where higher retention matters most: fewer derailments across multi-step tasks, fewer silent failures as the state evolves, and better consistency from one decision to the next. Here is Ternary Bonsai 2 27B running an agentic coding workflow with Cline on an NVIDIA https

1:19

131.9K views · 451 likes · 26 reposts · 9 replies Open on X →

PrismML@PrismML24 Jul

🦞🦞🦞 16.3K views · 54 likes · 3 reposts · 4 replies Open on X →

PrismML@PrismML24 Jul

♥️Huge thanks to @nvidia for the DGX Sparks for @openclaw engineering team. We have been putting these to hard work, product announcements soon with our joint efforts to improve local model experience and empowering use-cases locally beyond what is currently possible on https:/

46.7K views · 333 likes · 32 reposts · 24 replies Open on X →

PrismML@PrismML21 Jul

Try Ternary Bonsai 27B directly on Hugging Face , powered by @togethercompute https://t.co/yB7uIaHXpF https://t.co/PEDLAI2C6u

0:23

18.8K views · 129 likes · 15 reposts · 15 replies Open on X →

PrismML@PrismML14 Jul

@PrismML nice one!!! 5.3K views · 47 likes · 1 reposts · 1 replies Open on X →

PrismML@PrismML14 Jul

Links: Blog: https://t.co/GFFHOmzEf7 The Information: https://t.co/4xrfdMXoUH CNBC: https://t.co/kwV3QAeH4v Whitepaper: https://t.co/PwNgOpyuwV Models: https://t.co/6sbYg2vcMl WebGPU Demo: https://t.co/uNIoAZLFKt API: https://t.co/slXyzECdnc GitHub: https://t.co/3BAp6Rg6az Docs: 33.6K views · 319 likes · 33 reposts · 10 replies Open on X →

PrismML@PrismML14 Jul

The phone threshold is even harder than the storage number suggests. A phone exposes only part of its memory to an application, and the model must share that budget with its KV cache, activations, runtime, and the rest of the product. At 3.9 GB, 1-bit Bonsai 27B clears that htt

1:19

33.6K views · 219 likes · 14 reposts · 9 replies Open on X →

PrismML@PrismML14 Jul

Here is Ternary Bonsai 27B running an end-to-end agentic workflow locally with Hermes on an NVIDIA GeForce RTX 5090 GPU. The model reasons, calls tools, reads outputs, modifies files, and surfaces insights - all on consumer hardware, while all private files, intermediate states,

1:46

155.6K views · 547 likes · 55 reposts · 11 replies Open on X →

PrismML@PrismML14 Jul

The footprint reduction does not come at the expense of the capabilities that matter. Across 15 benchmarks spanning knowledge, reasoning, math, coding, instruction following, tool use, and vision, Ternary Bonsai 27B retains 95% of the full-precision model’s performance. The http

34.2K views · 272 likes · 22 reposts · 5 replies Open on X →

PrismML@PrismML14 Jul

Why does this matter? Because modern AI workflows are no longer single prompts. They are sustained loops. A capable agent may take hundreds of steps: reasoning, calling tools, reading outputs, updating its state, and iterating toward a result. When every step is remote, https://

27.4K views · 195 likes · 13 reposts · 1 replies Open on X →

PrismML@PrismML14 Jul

Today, we’re announcing Bonsai 27B: the first 27B-class model to run on a phone. Bonsai 27B is the new multimodal flagship of the Bonsai family. Based on Qwen3.6 27B, it brings a new capability tier to local AI: multi-step reasoning, structured tool use, long-context workflows,

1.5M views · 6K likes · 921 reposts · 309 replies Open on X →

PrismML@PrismML14 Jul

Raw capability determines what a model can do. Intelligence density determines where it can do it. Bonsai 27B moves the Pareto frontier left again: 27B-class capability in a footprint smaller than many full-precision 2B models. By intelligence density, 1-bit Bonsai 27B delivers

40.8K views · 300 likes · 14 reposts · 3 replies Open on X →

PrismML@PrismML26 May

Today we’re releasing 1-bit and Ternary Bonsai Image 4B. A new family of image-generation models designed to run high-quality diffusion inference on local hardware: from laptops to phones. https://t.co/9qB5UbOogJ

211.4K views · 1.7K likes · 235 reposts · 69 replies Open on X →

Aynı büyüklükteki hesaplara karşı

Son 90 güne ait 17 gönderi, 10K–100K takipçi aralığıyla yan yana. geniş kitleye gösteriliyor, ama izleyenlerin azı tepki veriyor.

Medyan görüntülenme40 809bu hesap99910K–100K için medyan

Erişim, %146.76%bu hesap3.88%10K–100K için medyan

Etkileşim, %0.77%bu hesap1.53%10K–100K için medyan

Ölçüt	Bu hesap	10K–100K için medyan	Oran
Gönderi başına medyan görüntülenme	40 809	999	40.9×
Erişim (görüntülenme ÷ takipçi)	146.76%	3.88%	37.8×
Etkileşim oranı	0.77%	1.53%	0.50×

Bu aralıktaki diğer hesaplar → Başka bir hesapla karşılaştır → Bu kıyas değerleri nasıl kuruluyor →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

34.2K14 Jul

155.6K

33.6K

5.3K

18.8K21 Jul

46.7K24 Jul

16.3K

131.9K17 Sep

183.4K

73.3K

111.6K

74K

19.1K

Last 14 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.89%14 Jul

0.40%

0.73%

1.09%

0.93%

0.86%21 Jul

0.84%24 Jul

0.37%

0.37%17 Sep

0.32%

0.77%

0.50%

0.31%

2.42%

Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.53%.

What the audience does

Likes56.1%12 753 in total

Reposts6.5%1 477 in total

Replies2.3%519 in total

Quotes3.1%703 in total

Bookmarks32.0%7 268 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

Followers by day

19 Sep

Daily snapshots since 19 Sep 2026; the dashed line is the starting count.

Son gönderiler

Aynı büyüklükteki hesaplara karşı

Growth & engagement

Views per post

Engagement rate per post

What the audience does

Followers by day

Benzer hesaplar