Unverified50% confidenceBenchmarkExact time
PrismML's Ternary Bonsai 2 applies ternary quantization to the Qwen3.8 27B model, has received 3.87 million downloads, compresses the language model from 54GB to 5.9GB — a 9.3x reduction — and officially claims it retains 98.2% of the FP16 model's capability, with an average score of 84.78 on 14 reasoning benchmarks.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/3/2026
First Seen
Valid until: 1/1/2027
Sources
Related Entities
Related Claims
UnverifiedBonsai 2, released by PrismML, is a heavily quantized version based on Qwen 3.8 27B.77% similarUnverifiedPrismML compressed Qwen 3.6 from the original approximately 54 GB to under 4 GB, a compression ratio exceeding 13×.73% similarUnverifiedPrismo claims its solution activates all 27 billion parameters simultaneously and compresses the original 54 GB QW3.6 model to less than 4 GB, a compression ratio of over 13x, with claimed no impact on performance.68% similarVerifiedQwen3-series coding models load at approximately 20 GB and can expand to 35–40 GB during inference.68% similarUnverifiedAfter quantization, the VRAM requirement of a 7-billion-parameter model is compressed from approximately 28 GB to approximately 4–6 GB, allowing consumer-grade GPUs such as the RTX 3060/4060 to run 7-billion-parameter-class models like Llama 3, Qwen2.5, and Mistral smoothly.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/970671API
curl https://kongchang.com/api/v1/knowledge/claims/970671MCP
get_claim(id=970671)