Ternary Bonsai 2 27B · PTQ1_0
This page aggregates 1 real-world runs of Ternary Bonsai 2 27B (PTQ1_0) on NVIDIA RTX 3080 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
52.18 tok/s
Decode
Decode speed
1,210.6 tok/s
Prefill
Prefill speed
— s
TTFT
Time to first token
7.03 GB
VRAM
VRAM usage
1
Measured runs
1
Independent sources
GitHub
Source platforms
5 days ago
Last verified
Performance
- llama.cpp · PTQ1_0 (current)Decode 52.18 · Prefill 1210.6 ·
Core figures
- Decode (avg)
- 52.18 tok/s
- Prefill (avg)
- 1,210.6 tok/s
- TTFT (avg)
- — s
- VRAM (avg)
- 7.03 GB
- MTP acceptance rate
- —
- TTFB
- — GB
- Power draw
- — W
Configuration
Member-level fields are taken from the most recent run
- Model
- Ternary Bonsai 2 27B
- Quantization
- PTQ1_0
- Framework
- llama.cpp
- Version
- prism-b10735-842b188
- Context length
- 32768 tokens
- Batch size
- 1
- GPU layers
- 99
- Flash Attention
- On
Hardware
Nominal and measured figures are shown side by side; whether it runs is the reader's call
- GPU
- NVIDIA RTX 3080
- Nominal VRAM
- 10 GB
- Measured VRAM (avg)
- 7.03 GB
- OS
- Ubuntu 24.04
- Driver
- 595.84
- CUDA
- 12.8
Get started
Original model resources
Base model resources may not match this quantization. Use the verified deployment weights above when available.
Startup command
Command from the latest benchmark run-ngl 99 -fa onSources & evidence
1 measured records in total, each traceable to its original source
prism-b10735-842b188 · Ubuntu 24.04 · CUDA 12.8 · 32768 ctx
52.18 tok/s
Decode
1,210.6 tok/s
Prefill
— s
TTFT
7.03 GB
VRAM
—
MTP
— W
Power draw
| pack | size | pp512 | tg128 | | PTQ1_0 (1.75 bpw) | 5.53 GiB | 1210.6 ± 29.8 | 52.18 ± 0.05 |