Skip to content
#33B+Jun 10, 2026Evidence B

ZDTaichu 5.0-9B

CASIA · 9B dense

A 9B Chinese multimodal veteran

A 9B multimodal update from CASIA in June 2026 under Apache-2.0: perception 88, Chinese 88, video 80 and a composite of 80.6. It has no especially sharp single dimension, but solid Chinese context and single-machine deployment make it a good private baseline for government and education. Ecosystem 78 is the weak spot.

TextImageVideoNative

Last updated Oct 21, 2026Editorial scoring

Bottom line

When you need a baseline with good Chinese context, private deployment and 9B practicality, ZDTaichu 5.0 is a safe option: Chinese 88, video 80, Apache-2.0. Do not expect it to shine in documents, agents or creative generation; it is a general base, not a single-dimension champion.

Eight-dimension breakdown

editorial mapping · 0–100
Perception88Document84Chart QA78Video80Agent70Generation74Chinese88Ecosystem788D
  • ZDTaichu 5.0-9B80.6
Visual perception & grounding20%
88
Document parsing & Chinese OCR18%
84
Chart & document reasoning14%
78
Video & long-horizon reasoning14%
80
GUI agents & tool use14%
70
Generation & editing6%
74
Chinese context & culture6%
88
Ecosystem, cost & availability8%
78

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9BDense model
许可证Apache-2.0Commercial use
上下文128KOfficial 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 OCR84 · Fine on clean receipts
Chart data extraction78 · Stable on simple charts
30-min video QA80 · Above-average event localization
Mobile GUI control68 · Not GUI-specific
Chinese cultural context88 · Solid Chinese context

Strengths and limitations

Strengths

  • Chinese 88 and video 80 are a rare balanced combination at 9B.
  • Apache-2.0 plus 9B keeps the barrier low for government and education on-prem work.
  • Perception 88 keeps real-world images stable.

Limitations

  • Documents 84 and charts 78: complex layouts and charts are not strengths.
  • Agents 70: complex orchestration will not hold.
  • Ecosystem 78: tooling and community reproduction trail the internet giants.

Price and real cost

List price

—N/A

Open weights; self-hosted, no API fee. For 1,000 images at 1,024px, ~1,100 image tokens each and 300 output tokens, the monthly cost is about ¥0.0 (assuming input at 25% of output price).

Use your volume

About ZDTaichu 5.0-9B

Is ZDTaichu 5.0-9B open source?

Licence: Open weights (Apache-2.0). Weights are available, but check the exact terms and filing requirements before commercial use.

How much does ZDTaichu 5.0-9B cost?

Open weights; self-hosted, no API fee. It ranks #33 with a composite of 80.6. Real bills depend on resolution, tiling, caching and volume discounts.

What is ZDTaichu 5.0-9B best for?

Government and education on-prem; Chinese content understanding; Video triage。

Why does ZDTaichu 5.0-9B score 80.6?

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 Visual perception & grounding; its weakest is GUI agents & tool use.

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.