Back to search results
Qwen3.5-0.8B · Q8_0
NVIDIA RTX 507012GBllama.cpp b4200L0 Self-reported
This page aggregates 1 real-world runs of Qwen3.5-0.8B (Q8_0) on NVIDIA RTX 5070 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
582.4 tok/s
Decode
Decode speed
4,368 tok/s
Prefill
Prefill speed
0.02 s
TTFT
Time to first token
1.4 GB
VRAM
VRAM usage
L0 Self-reported
1
Measured runs
1
Independent sources
Other
Source platforms
2026-10-06
Last verified
Performance
- llama.cpp · Q4_K_M Decode 186.5 · Prefill 627.46 ·
- Ollama · FP16 Decode 84.08 · Prefill 290.65 ·
- Ollama · IQ1_S Decode 239.86 · Prefill 1174.67 ·
- vLLM · IQ3_XS Decode 223.3 · Prefill 1047.53 ·
- vLLM · IQ4_XS Decode 188.25 · Prefill 873.82 ·
- Ollama · Q3_K_S Decode 197.53 · Prefill 685.98 ·
- Ollama · Q5_K_S Decode 124.72 · Prefill 677.51 ·
- Strata · IQ4_XS Decode 233.02 · Prefill 915.89 ·
- Strata · Q6_K Decode 189.01 · Prefill 746.35 ·
- llama.cpp · Q8_0 (current)Decode 582.4 · Prefill 4368 ·
Core figures
- Decode (avg)
- 582.4 tok/s
- Prefill (avg)
- 4,368 tok/s
- TTFT (avg)
- 0.02 s
- VRAM (avg)
- 1.4 GB
- MTP acceptance rate
- —
- TTFB
- — GB
- Power draw
- — W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen3.5-0.8B
- Quantization
- Q8_0
- Framework
- llama.cpp
- Version
- b4200
- Context length
- 4096 tokens
- Batch size
- 512
- 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 5070
- Nominal VRAM
- 12 GB
- Measured VRAM (avg)
- 1.4 GB
- OS
- Linux
- Driver
- 570.65
- CUDA
- 12.6
Sources & evidence
1 measured records in total, each traceable to its original source
b4200 · Linux · CUDA 12.6 · 4096 ctx
582.4 tok/s
Decode
4,368 tok/s
Prefill
0.02 s
TTFT
1.4 GB
VRAM
—
MTP
— W
Power draw
L0 理论估算:基于显存带宽 ÷ 模型体积公式,结合框架实测效率因子校准