Arao AI
Back to search results

Qwen3.5-0.8B · Q3_K_L

Start a discussion
NVIDIA RTX 3060 12GB12GBOllama v0.50L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-0.8B (Q3_K_L) on NVIDIA RTX 3060 12GB with Ollama, contributed by 1 independent source platforms; metrics are averages of published measurements.

167.74 tok/s

Decode

Decode speed

790.34 tok/s

Prefill

Prefill speed

0.62 s

TTFT

Time to first token

3.31 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

Today

Last verified

Performance

  1. llama.cpp · Q3_K_S Decode 187.82 · Prefill 792.93 ·
  2. llama.cpp · Q5_K_S Decode 142.06 · Prefill 686.26 ·
  3. Ollama · Q3_K_L (current)Decode 167.74 · Prefill 790.34 ·
  4. Ollama · Q5_K_M Decode 118.62 · Prefill 422.11 ·
  5. Strata · Q3_K_M Decode 279.25 · Prefill 1294.74 ·
  6. Strata · Q5_K_S Decode 214.78 · Prefill 1075.84 ·
  7. llama.cpp · Q8_0 Decode 312 · Prefill 2340 ·

Core figures

Decode (avg)
167.74 tok/s
Prefill (avg)
790.34 tok/s
TTFT (avg)
0.62 s
VRAM (avg)
3.31 GB
MTP acceptance rate
—
TTFB
1.34 GB
Power draw
162.9 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-0.8B
Quantization
Q3_K_L
Framework
Ollama
Version
v0.50
Context length
16384 tokens
Batch size
512
GPU layers
20
Flash Attention
Off

Hardware

Nominal and measured figures are shown side by side; whether it runs is the reader's call

GPU
NVIDIA RTX 3060 12GB
Nominal VRAM
12 GB
Measured VRAM (avg)
3.31 GB
OS
macOS 15.0
Driver
560.94
CUDA
12.4
Power draw
162.9 W

Sources & evidence

1 measured records in total, each traceable to its original source

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-10

    v0.50 · macOS 15.0 · CUDA 12.4 · 16384 ctx

    167.74 tok/s

    Decode

    790.34 tok/s

    Prefill

    0.62 s

    TTFT

    3.31 GB

    VRAM

    —

    MTP

    162.9 W

    Power draw

    L2 batch seed 2026-10-10: Ollama on hw_id=15 running model_id=51 at Q3_K_L