| Setup | Best plan | Memory | Generate |
|---|---|---|---|
| 8 GB GPU + 32 GB RAM | Ternary-Bonsai-2-9B-Q2_0llama.cpp CUDAFull GPU | KV unknown | Unknown |
| 12 GB GPU + 32 GB RAM | Ternary-Bonsai-2-9B-Q2_0llama.cpp CUDAFull GPU | KV unknown | Unknown |
| 16 GB GPU + 64 GB RAM | Ternary-Bonsai-2-9B-Q2_0llama.cpp CUDAFull GPU | KV unknown | Unknown |
| 24 GB GPU + 64 GB RAM | Ternary-Bonsai-2-9B-Q2_0llama.cpp CUDAFull GPU | KV unknown | Unknown |
| 32 GB GPU + 128 GB RAM | Ternary-Bonsai-2-9B-Q2_0llama.cpp CUDAFull GPU | KV unknown | Unknown |
| 2× 24 GB GPUs + 128 GB RAM | Ternary-Bonsai-2-9B-Q2_0llama.cpp CUDAFull GPU | KV unknown | Unknown |
| 96 GB workstation GPU + 256 GB RAM | Ternary-Bonsai-2-9B-Q2_0llama.cpp CUDAFull GPU | KV unknown | Unknown |
| Mac, 24 GB | Ternary-Bonsai-2-9B-Q2_0llama.cpp MetalUnified memory | KV unknown | Unknown |
| Mac, 48 GB | Ternary-Bonsai-2-9B-Q2_0llama.cpp MetalUnified memory | KV unknown | Unknown |
| Mac, 128 GB | Ternary-Bonsai-2-9B-Q2_0llama.cpp MetalUnified memory | KV unknown | Unknown |
| Mac, 512 GB | Ternary-Bonsai-2-9B-Q2_0llama.cpp MetalUnified memory | KV unknown | Unknown |
| DGX Spark 128 GB | Ternary-Bonsai-2-9B-Q2_0llama.cpp CUDAUnified memory | KV unknown | Unknown |
| Ryzen AI Max+ 395, 128 GB | Ternary-Bonsai-2-9B-Q2_0llama.cpp VulkanUnified memory | KV unknown | Unknown |
| CPU only, 64 GB DDR5-5600 dual channel | Ternary-Bonsai-2-9B-PQ2_0llama.cpp CPUCPU only | KV unknown | Unknown |
Same plan engine as the planner, with its default reserves and RAM. How plans are chosen.
| Artifact | Size | Files | Revision | Source |
|---|---|---|---|---|
| Ternary-Bonsai-2-9B-PQ2_0 native GGUF option prism-ml/Ternary-Bonsai-2-9B-gguf | 2.25 GiB | 1 | 5b2abb0236 | HF files |
| Ternary-Bonsai-2-9B-PTQ1_0 native GGUF option prism-ml/Ternary-Bonsai-2-9B-gguf | 1.86 GiB | 1 | 5b2abb0236 | HF files |
| Ternary-Bonsai-2-9B-Q2_0 native GGUF option prism-ml/Ternary-Bonsai-2-9B-gguf | 2.38 GiB | 1 | 5b2abb0236 | HF files |
Not found in the representative quant search (not proven unavailable): GGUF FP16, GGUF Q8_0, GGUF Q5_K_M, GGUF Q4_K_M, MLX 8-bit, MLX 4-bit, EXL3, AWQ, GPTQ. Complete selected file sets; projectors, MTP/draft and lookup files are listed separately and budgeted only by plans that use them. File bytes are not allocated memory.
No Artificial Analysis evaluation mapped to this checkpoint.
Arena text rating
Unknown
Publisher-reported benchmarks
None extracted from the model card.
AA and Arena describe the hosted model in the stated mode, not each quantized file; scores are never inherited by base, pruned, abliterated or quantized variants and never averaged together.
PrismML llama.cpp fork
source-documented, not executed source · prism-b10776-d00e252 orlater forMTP+quantizedKV onCUDA;branchprism
./bin/llama-server -m Ternary-Bonsai-2-9B-PQ2_0-MTP-Q8_0.gguf -ngl 99 -fa on \ --spec-type draft-mtp --spec-draft-n-max 2 --spec-draft-n-min 0 --spec-draft-p-min 0
Recipes are source examples (settings are the publisher’s or author’s, not measured optima) and were not executed by LocalFit. A downloadable file does not prove your backend build supports the architecture.
No reported measurement yet. Submit one.