Skip to content
Benchmarks / Vision /Frontier / BF16

Vision · Frontier · BF16

A 4B-class model at BF16 on 12 GB. The universal baseline: every supported NVIDIA GPU from an RTX 3060 12 GB upward runs this identically, so it is the one leaderboard that spans the whole hardware range.

Best Images / min
img/min
Ranked results
0
0 distinct systems
GPU models
0
Minimum VRAM
64 GB
~14 min run

Best result per GPU

Images / min · img/min
No results yet for this view.
Show the values behind this chart (0 rows)
GPUGPUsSamplesBest (img/min)Mean (img/min)

Download these values as CSV· Full open dataset

No results match these filters yet.

What is pinned

Everything below is fixed by the profile and checked on every submission. A result whose runtime flags differ from these is published, but never ranked.

Model
Repository
Qwen/Qwen3-VL-32B-Instruct
Revision
pending freeze
Precision
bf16
Parameters
32 B
Licence
Apache-2.0
Runtime
Engine
vllm 0.26.0
Harness
aiperf
dtype
bfloat16
max-model-len
32768
gpu-memory-utilization
0.9
max-num-seqs
16
enforce-eager
false
disable-log-requests
true
tensor-parallel-size
1
swap-space
0
Workloads
vision_ocr
task=ocr image_long_edge_px=1080 max_output_tokens=256 concurrency=1 temperature=0
vision_chart
task=chart image_long_edge_px=1080 max_output_tokens=256 concurrency=1 temperature=0
vision_document
task=document image_long_edge_px=1080 max_output_tokens=256 concurrency=1 temperature=0
vision_description
task=description image_long_edge_px=1080 max_output_tokens=256 concurrency=1 temperature=0
vision_reasoning
task=reasoning image_long_edge_px=1080 max_output_tokens=256 concurrency=1 temperature=0
vision_throughput
task=description image_long_edge_px=1080 max_output_tokens=128 concurrency=2 temperature=0
Ranking
Ranked on
Images / min (img/min, higher is better)
Gate
Answer accuracy ≥ 90
Gate
Error rate ≤ 0
Also shown
TTFT p50, TTFT p95, End to end p50, Decode, Image encode p50, Energy / image
Design notes
  • Release candidate. The runtime and harness versions are pinned; the model revision is not frozen until the bake-off confirms fit on the tier's minimum-VRAM reference system, so this leaderboard is provisional.
  • Headline is images per minute at concurrency 2, which is what a reader actually wants to know.
  • Quality is a gate, not a score. OCR, chart and document tasks have deterministic expected answers; description and reasoning are collected but never scored until a stable evaluator exists.
  • Image encode time is separated from prompt prefill so the vision tower cost is visible.