// Get open-weight models
Ternary Bonsai 2 27B
Prism ML’s ternary (−1/0/+1) build of Qwen3.8-27B: ~1.72 bits per language weight, GGUF ~5.95–7.21GB, the card claims 98.2% of FP16 quality and 262K context; runs on its llama.cpp fork (CUDA / Metal / CPU), with an MLX build for Apple silicon. Apache-2.0
Open source Maintained For developers
IN ONE PARAGRAPH
Ternary Bonsai 2 27B is prism ML’s ternary (−1/0/+1) build of Qwen3.8-27B: ~1.72 bits per language weight, GGUF ~5.95–7.21GB, the card claims 98.2% of FP16 quality and 262K context; runs on its llama.cpp fork (CUDA / Metal / CPU), with an MLX build for Apple silicon. Apache-2.0.
Good for: Developers and teams who want to get open-weight models.
01How to start
- 1Read the official quick-start (link above).
- 2Install or sign up as the official guide describes.
- 3Try it on a small example before wiring it into a project.
02Alternatives & comparison
Compare side by side → DeepSeek-V3 Open-weight models ★ 105k Qwen3 Qwen3 is the large language model series developed by Qwen team, Alibaba Cloud. ★ 28k gpt-oss gpt-oss-120b and gpt-oss-20b are two open-weight language models by OpenAI ★ 20k Qwen3-Coder Qwen3-Coder is the code version of Qwen3, the large language model series developed by Qwen team. ★ 17k Kimi-K2 Kimi K2 is the large language model series developed by Moonshot AI team ★ 11k
Versions, why it is listed, for AI to read
For AI agents
- Interfaces：its own app or website
- Machine-readable record：/api/v1/software/ternary-bonsai-2-27b.json · llms.txt
- Added 2026-10-20