Ternary Bonsai 2 27B
PrismML · Released Sep 2026
Conditional
Ternary Bonsai 2 27B squeezes Qwen3.8 27B into about 5.9GB at 1.76 effective bits, so a 262K-context text-and-image model fits on a Mac or a single consumer GPU under Apache 2.0. Conditional because the quality figures are PrismML's own, its blog and model card report different aggregate scores, and the files need PrismML's custom kernels: stock llama.cpp will not run them. Right for local-inference users short on memory.
Is it right for you?
Good for
- Running a 27B-class model in about 6GB: the ternary GGUF is 5.95GB against a 53.8GB FP16 reference, plus a 0.63GB projector for image input
- Apple silicon Macs: PrismML reports 46.8 tokens per second on an M5 Max through MLX
- Long-context local work with a 262K-token window
- Commercial use and redistribution under Apache 2.0
Not good for
- Standard local tooling: the GGUF files need PrismML's llama.cpp fork, and stock llama.cpp will not run them
- Maximum quality: PrismML's own figures put it at 98.2% of full-precision Qwen3.8 27B
- Decisions that need independent evaluations: all published scores are PrismML's
Pricing
Input
Free (open weights)
Output
Free (open weights)
Context
262K ctx; self-hosted; PrismML kernels required
Benchmarks
No verdict changes yet
The clock starts day one. Changes land here as our verdict evolves.
Sources
Verification log
- Pricing· No changes
Automated agent
- Profile· No changes
Automated agent
- Pricing· No changes
Automated agent
- Profile· No changes
Automated agent
- Pricing· No changes
Automated agent