Back to search results
Qwen3.5-4B · Q4_K_M
NVIDIA RTX 507012GBvLLM v0.22L2 Cross-framework verified
This page aggregates 1 real-world runs of Qwen3.5-4B (Q4_K_M) on NVIDIA RTX 5070 with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.
84.51 tok/s
Decode
Decode speed
408.98 tok/s
Prefill
Prefill speed
1.63 s
TTFT
Time to first token
8.99 GB
VRAM
VRAM usage
L2 Cross-framework verified
1
Measured runs
1
Independent sources
Other
Source platforms
Today
Last verified
Performance
- vLLM · Q3_K_M Decode 112.72 · Prefill 436.5 ·
- vLLM · Q4_K_M (current)Decode 84.51 · Prefill 408.98 ·
- Strata · IQ3_S Decode 154.26 · Prefill 563.04 ·
- llama.cpp · Q5_K_M Decode 174.7 · Prefill 1310.3 ·
Core figures
- Decode (avg)
- 84.51 tok/s
- Prefill (avg)
- 408.98 tok/s
- TTFT (avg)
- 1.63 s
- VRAM (avg)
- 8.99 GB
- MTP acceptance rate
- —
- TTFB
- 0.68 GB
- Power draw
- 155.6 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen3.5-4B
- Quantization
- Q4_K_M
- Framework
- vLLM
- Version
- v0.22
- Context length
- 4096 tokens
- Batch size
- 16
- GPU layers
- 999
- Flash Attention
- Off
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)
- 8.99 GB
- OS
- Ubuntu 22.04
- Driver
- 565.57
- CUDA
- 12.8
- Power draw
- 155.6 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.22 · Ubuntu 22.04 · CUDA 12.8 · 4096 ctx
84.51 tok/s
Decode
408.98 tok/s
Prefill
1.63 s
TTFT
8.99 GB
VRAM
—
MTP
155.6 W
Power draw
L2 batch seed 2026-10-10: vLLM on hw_id=18 running model_id=48 at Q4_K_M