Skip to content
#22ASep 5, 2026Evidence B

Ling-3.0-flash-VL

Ant Group · 124B total / 5.5B active (MoE)

A lightweight open tier for Chinese documents and finance

A 124B-A5.5B vision-language model released by Ant in September 2026, MIT and 256K: documents 90, Chinese 90, perception 89 and a composite of 85.9. It handles text and images only; video 84 means “watchable but not precise”, and generation is 72. Its advantages are low cost, strong Chinese context and clean licensing.

TextImageNative

Last updated Oct 21, 2026Editorial scoring

Bottom line

A good fit for Chinese document pipelines in finance and government: documents 90, Chinese 90, MIT, ¥0.8 — the batch economics work. Do not use it for video, creative generation or complex agents; it is a cheap and reliable Chinese image-text component.

Eight-dimension breakdown

editorial mapping · 0–100
Perception89Document90Chart QA85Video84Agent82Generation72Chinese90Ecosystem888D
  • Ling-3.0-flash-VL85.9
Visual perception & grounding20%
89
Document parsing & Chinese OCR18%
90
Chart & document reasoning14%
85
Video & long-horizon reasoning14%
84
GUI agents & tool use14%
82
Generation & editing6%
72
Chinese context & culture6%
90
Ecosystem, cost & availability8%
88

Weights: perception 20 / documents 18 / charts 14 / video 14 / agents 14 / generation 6 / Chinese 6 / ecosystem 8. Definitions and sources on the methodology page.

Public benchmarks

as reported, not converted
Benchmark / metricValueNote
参数规模124B-A5.5BOfficial spec
许可证MITCommercial use
上下文256KOfficial spec

Public scores come from different versions and prompts and are not interchangeable with our composite; they are one evidence input.

Scenario mapping

five stress tests, one scale
ScenarioScoreObservation
Chinese receipt OCR90 · Stable on clean receipts
Chart data extraction85 · Fine on standard charts
30-min video QA50 · No video support
Mobile GUI control78 · Fine for simple tasks
Chinese cultural context92 · Chinese context is the strength

Strengths and limitations

Strengths

  • Chinese 90 and documents 90 keep official Chinese documents and receipts stable.
  • MIT plus ¥0.8 suits finance and government batch work on both licence and cost.
  • 256K context reduces chunking for long documents.

Limitations

  • Text and images only; no video or audio.
  • Generation 72: no image generation or editing.
  • Agents 82 and ecosystem 88: complex orchestration still needs an upper-layer framework.

Price and real cost

List price

¥0.8/M tokens

≈ ¥0.8 / 1M output tokens; weights free. For 1,000 images at 1,024px, ~1,100 image tokens each and 300 output tokens, the monthly cost is about ¥0.5 (assuming input at 25% of output price).

Use your volume

About Ling-3.0-flash-VL

Is Ling-3.0-flash-VL open source?

Licence: Open weights (MIT). Weights are available, but check the exact terms and filing requirements before commercial use.

How much does Ling-3.0-flash-VL cost?

≈ ¥0.8 / 1M output tokens; weights free. It ranks #22 with a composite of 85.9. Real bills depend on resolution, tiling, caching and volume discounts.

What is Ling-3.0-flash-VL best for?

Chinese official documents and contracts; Financial receipt batches; On-prem image-text work under MIT。

Why does Ling-3.0-flash-VL score 85.9?

The composite weights eight dimensions: perception 20%, documents 18%, charts 14%, video 14%, agents 14%, generation 6%, Chinese 6%, ecosystem 8%. Its strongest dimension is Document parsing & Chinese OCR; its weakest is Generation & editing.

Personal ranking · Not an official score

Every ranking, score and comparison on this page is an independent personal ranking, not an official score, and represents no vendor, institution or regulator. Scores are a subjective mapping of public evidence under our own weights — change the weights and the order changes. Disagree, or spotted a factual error? Contact the site owner: jizhi2024@foxmail.com — we check each report and log corrections in the changelog.