Arao AI
Back to search results

Qwen3-Coder-Next · FP16

Start a discussion
Apple M1Ollama v0.21L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-Coder-Next (FP16) on Apple M1 with Ollama, contributed by 1 independent source platforms; metrics are averages of published measurements.

21.35 tok/s

Decode

Decode speed

72.87 tok/s

Prefill

Prefill speed

0.51 s

TTFT

Time to first token

1.05 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

2 days ago

Last verified

Performance

  1. vLLM · IQ4_XS Decode 47.87 · Prefill 181.96 ·
  2. Ollama · FP16 (current)Decode 21.35 · Prefill 72.87 ·
  3. Ollama · Q3_K_L Decode 41.49 · Prefill 134.08 ·

Core figures

Decode (avg)
21.35 tok/s
Prefill (avg)
72.87 tok/s
TTFT (avg)
0.51 s
VRAM (avg)
1.05 GB
MTP acceptance rate
—
TTFB
0.17 GB
Power draw
69.5 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-Coder-Next
Quantization
FP16
Framework
Ollama
Version
v0.21
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
Apple M1
Measured VRAM (avg)
1.05 GB
OS
macOS 15.0
Driver
Apple M1
CUDA
N/A
Power draw
69.5 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-16

    v0.21 · macOS 15.0 · CUDA N/A · 16384 ctx

    21.35 tok/s

    Decode

    72.87 tok/s

    Prefill

    0.51 s

    TTFT

    1.05 GB

    VRAM

    —

    MTP

    69.5 W

    Power draw

    L2 batch seed 2026-10-16: Ollama on hw_id=99 running model_id=58 at FP16