Editorial intelligence brief
What matters now
Released the Apache-2.0 Ternary Bonsai 2 27B on September 17, 2026, a 5.9 GB multimodal model based on Qwen3.8 27B with a 262K-token context window.
Focus: Low-bit multimodal foundation models for efficient local, edge, and datacenter deployment
Latest funding
Not tracked
No structured event
Tracked funding
Not tracked
No structured events
Latest valuation
Not tracked
No source-backed record
Latest revenue
Not tracked
Deployment lifecycle position
Source-backed financial history
Funding and financial history
Funding, valuation, revenue, and debt are kept separate so unlike measurements are never presented as one trend.
A structured financial history has not been completed for PrismML.
The profile will not manufacture a trend from incomplete or unsourced figures. Each amount must be dated, typed, and linked to public evidence first.
Chronological activity
What changed
Released the Apache-2.0 Ternary Bonsai 2 27B on September 17, 2026, a 5.9 GB multimodal model based on Qwen3.8 27B with a 262K-token context window.
A structured event record is still being assembled for this company.
Recent coverage of PrismML
Consolidated by storyPrismML Releases Ternary Bonsai 2 27B: A 5.9 GB Apache 2.0 Model Retaining 98.2% of Qwen3.8 27B Performance
marktechpost.com · 1 related source · Sep 18
PrismML launches Bonsai 2 27B, a high-intelligence AI model so small it fits on consumer hardware
siliconangle.com · Sep 18
Deploying a 1-Bit Bonsai-27B Model with PrismML llama.cpp and OpenAI-Compatible Local Inference Workflows
marktechpost.com · Jul 28
Product surface
Products and model families
Recent models
Products
Open weights
Yes
API access
No
Review cadence
Weekly
Market position
Lifecycle and strategic signals
Research
Preview
Growth
Deployment
Deployment: Production & at scale
Focus
Low-bit multimodal foundation models for efficient local, edge, and datacenter deployment
Signals
Taxonomy
Competitive context
Infrastructure & distribution
Where PrismML sits in the AI stack
Model layer
4 tracked model families
Access layer
Self-hostable weights
Market layer
Open Weight Labs
Public product surface
Stack signals
This section reflects publicly visible product and access signals. Undisclosed GPU, cloud, training, and data-center contracts are not inferred.
Evidence trail
Sources and corrections
Primary and public sources used to verify the structured company profile.
The Staticbreaker
Get the signal behind the week's AI changes.
The Staticbreaker turns launches, category shifts, and market noise into a concise editorial briefing.