{"apiVersion":"v1","taxonomyVersion":1,"generatedAt":"2026-10-29T13:28:24.318Z","data":{"recordId":"wsa_2k9kkmugid49","score":0.938,"confidence":"high","capped":false,"capReason":null,"sources":[{"n":1,"url":"https://huggingface.co/prism-ml/Ternary-Bonsai-2-4B-gguf","publisher":"PrismML","authority":"owner-controlled","verification":"verified","lastRetrievedAt":"2026-10-28T00:02:38.647Z","detectedPublishedAt":"2026-09-23T06:04:18.000Z","rights":"original"},{"n":2,"url":"https://huggingface.co/prism-ml/Ternary-Bonsai-2-4B-mlx-2bit","publisher":"PrismML","authority":"owner-controlled","verification":"verified","lastRetrievedAt":"2026-10-28T00:02:38.647Z","detectedPublishedAt":"2026-10-01T17:06:34.000Z","rights":"original"},{"n":3,"url":"https://prismml.com/news/bonsai-2-family","publisher":"PrismML","authority":"owner-controlled","verification":"verified","lastRetrievedAt":"2026-10-28T00:02:38.647Z","detectedPublishedAt":"2026-10-15T00:00:00.000Z","rights":"original"},{"n":4,"url":"https://prismml.com/news/prismml-expands-bonsai-2-family-with-2b-4b-and-9b-models-for-on-device-ai","publisher":"PrismML","authority":"owner-controlled","verification":"verified","lastRetrievedAt":"2026-10-28T00:02:38.647Z","detectedPublishedAt":"2026-10-15T00:00:00.000Z","rights":"original"}],"claims":[{"id":"29d06fd4-6f70-4f6b-b83e-57173ceeccdb","kind":"capability","statement":"The model supports reasoning, visual understanding, and tool-driven workflows.","citations":[{"source":3,"supportProbability":0.71,"supports":true,"quote":null},{"source":4,"supportProbability":0.62,"supports":false,"quote":{"text":"Building on the first-generation Bonsai 1.7B, 4B and 8B models, the new releases combine improved reasoning, visual understanding, and support for tool-driven workflows.","rights":"original"}}]},{"id":"2d72a460-79ad-4f44-9358-121c70847107","kind":"availability","statement":"The model is downloadable starting October 15, 2026.","citations":[{"source":4,"supportProbability":0.8,"supports":true,"quote":{"text":"Developers, researchers, and other users can download the new Bonsai 2 models under the Apache 2.0 license for free starting October 15, 2026.","rights":"original"}}]},{"id":"69d22d2d-eeb0-4859-a1d6-fb4b34efd3ff","kind":"deployment","statement":"The model is optimized for low-latency inference on consumer-grade CPUs and edge GPUs.","citations":[{"source":4,"supportProbability":0.8,"supports":true,"quote":{"text":"They have been designed for seamless integration with existing AI workflows and are optimized for low-latency inference on consumer-grade CPUs, and edge GPUs.","rights":"original"}}]},{"id":"53067df6-f27e-4338-b5aa-400e5d95a41a","kind":"recommendation-relevant","statement":"The model has a 0.95GB footprint.","citations":[{"source":3,"supportProbability":0.98,"supports":true,"quote":null},{"source":4,"supportProbability":0.96,"supports":true,"quote":{"text":"The models use Prism ML ’s ternary weight representation, which encodes model weights using three values: {−1, 0, +1}.","rights":"original"}}]},{"id":"a747a8cc-2d9a-4e54-ad37-d4d722864d24","kind":"capability","statement":"The model accepts vision input through the official vision tower.","citations":[{"source":2,"supportProbability":0.97,"supports":true,"quote":{"text":"- **Vision input** through the official vision tower, carried unrotated in FP16.","rights":"original"}}]},{"id":"a8360824-76c1-42e3-a7a0-819e67929662","kind":"deployment","statement":"The model runs with MLX on Apple Silicon (Metal).","citations":[{"source":2,"supportProbability":0.94,"supports":true,"quote":{"text":"| Backends | MLX on Apple Silicon (Metal): mlx-vlm 0.7.8+ natively, or the loader in `runtime/` |","rights":"original"}}]},{"id":"e4300809-860b-4b11-86b5-32d29a5ab2cb","kind":"capability","statement":"The optional multi-token-prediction head is intended to help most on copy-heavy work such as code edits and rewrites.","citations":[{"source":2,"supportProbability":0.97,"supports":true,"quote":{"text":"It helps most on copy-heavy work such as code edits and rewrites; on general chat it is about as fast as plain decoding or slightly slower, so plain decoding stays the default.","rights":"original"}}]},{"id":"1ec23cd1-8410-4975-b0aa-b5fb2e640c4a","kind":"capability","statement":"The model supports vision input through a separate mmproj pack.","citations":[{"source":1,"supportProbability":0.99,"supports":true,"quote":{"text":"- **Vision input** through a separate mmproj pack (Q8_0, or BF16 for reference precision), loaded only when an image arrives.","rights":"original"}}]},{"id":"d58c640c-a477-48ac-b08d-cef0a8bea54a","kind":"capability","statement":"The model has a 262,144-token context length inherited from its base model.","citations":[{"source":1,"supportProbability":0.98,"supports":true,"quote":{"text":"| Context length | 262,144 tokens (inherited from the base model) |","rights":"original"}},{"source":2,"supportProbability":0.81,"supports":true,"quote":null}]},{"id":"e72e818d-d431-46ef-8a0a-4666f2bf22fe","kind":"availability","statement":"Three GGUF packings are available: PTQ1_0, PQ2_0, and Q2_0.","citations":[{"source":1,"supportProbability":0.98,"supports":true,"quote":{"text":"- **Three GGUF packings** of the same weights, consumed directly by custom kernels and never expanded back to FP16: **PTQ1_0** (dense trits, 0.95 GB), **PQ2_0** (2-bit slots, 1.15 GB) and **Q2_0** (1.21 GB).","rights":"original"}}]},{"id":"88542a0d-9e6e-48cc-904e-0bf55d8934a8","kind":"deployment","statement":"The PrismML llama.cpp fork runs every listed GGUF file on CUDA, Metal, and CPU.","citations":[{"source":1,"supportProbability":0.55,"supports":false,"quote":{"text":"| Backends | llama.cpp (CUDA, Metal, CPU) via the PrismML fork |","rights":"original"}}]},{"id":"6d8ee70f-4d14-4f33-a583-32183aaa9a70","kind":"deployment","statement":"Upstream llama.cpp runs only the Q2_0 packing.","citations":[{"source":1,"supportProbability":0.97,"supports":true,"quote":{"text":"Upstream llama.cpp runs only `Q2_0` (recent `master`, which includes the Hadamard activation runtime); it rejects `PQ2_0` and `PTQ1_0` as unknown types.","rights":"original"}}]},{"id":"ae68bb37-eead-47c6-9494-a421ba9abce2","kind":"availability","statement":"The model is licensed under Apache 2.0.","citations":[{"source":1,"supportProbability":0.99,"supports":true,"quote":null},{"source":2,"supportProbability":0.98,"supports":true,"quote":null}]}]},"nextCursor":null,"license":{"catalog":{"name":"CC BY 4.0","spdx":"CC-BY-4.0","url":"https://creativecommons.org/licenses/by/4.0/","attribution":"wowsuchai","attributionUrl":"https://wowsuchai.withglint.app","covers":["catalog metadata","classifications","generated summaries","feed event metadata"]},"excluded":{"covers":["source quotations","linked source content"],"terms":"Not licensed by wowsuchai. Source quotations and linked source content keep their original rights."}}}