Head-to-head
Step-5-Pre vs Step-3
A 8.3-point composite gap — but the real story is the dimension mix. Same weights, unpacked: who leads where, and whether the premium is justified.
The short answer
The composites differ by 8.3 points, and their strengths do not overlap.
Stronger overall: Step-5-Pre. If you only run one workload, read the dimension table instead.
Overlaid radar
Same weights, same scale; overlap means parity
- Step-5-Pre93.0
- Step-384.7
Dimension deltas
A on the left, B on the right; 100 max per row
Step-5-Pre leads 7, Step-3 leads 1.
Spec sheet
Hard specs in one table
| Item | Step-5-Pre | Step-3 |
|---|---|---|
| Composite | 93.0 | 84.7 |
| Tier | S | A- |
| Visual perception & grounding | 96WIN | 88 |
| Document parsing & Chinese OCR | 95WIN | 85 |
| Chart & document reasoning | 93WIN | 84 |
| Video & long-horizon reasoning | 91WIN | 86 |
| GUI agents & tool use | 94WIN | 85 |
| Generation & editing | 78 | 80WIN |
| Chinese context & culture | 93WIN | 85 |
| Ecosystem, cost & availability | 94WIN | 78 |
| Output price | ¥4 / M | ¥5 / M |
| Context | 128K | 256K |
| Parameters | Undisclosed (MoE) | Undisclosed (MoE) |
| Licence | Commercial + on-prem licence | Closed commercial |
| Native multimodal | Yes | Yes |
| Evidence grade | B | B |
How to choose
Choose Step-5-Pre if…
- A flagship-class stand-in for edge-cloud products
- Production video understanding
- Speech-plus-vision-plus-video devices
- You cannot accept its limitation: Generation and editing at 78 — effectively no usable text-to-image or Chinese glyph rendering.
How to choose
Choose Step-3 if…
- Edge-cloud products
- In-car multimodal interaction
- Light speech-plus-vision apps
- You cannot accept its limitation: Fewer public evals and docs.
Scenario check
Five scenarios, who holds up
| Scenario | Step-5-Pre | Step-3 | Better |
|---|---|---|---|
| Chinese receipt OCR | 95 | 85 | Step-5-Pre |
| Chart data extraction | 93 | 83 | Step-5-Pre |
| 30-min video QA | 91 | 86 | Step-5-Pre |
| Mobile GUI control | 94 | 85 | Step-5-Pre |
| Chinese cultural context | 93 | 84 | Step-5-Pre |
Comparison FAQ
Step-5-Pre vs Step-3: which is stronger?
Composite: Step-5-Pre 93.0, Step-3 84.7. Step-5-Pre leads on Visual perception & grounding, Document parsing & Chinese OCR, Chart & document reasoning, Video & long-horizon reasoning, GUI agents & tool use, Chinese context & culture, Ecosystem, cost & availability; Step-3 leads on Generation & editing. The stronger overall is Step-5-Pre.
How much more expensive is Step-5-Pre than Step-3?
Step-5-Pre is ¥4/1M output tokens; Step-3 is ¥5/1M. On output price, Step-3 is pricier; real cost also depends on image tokens and input share.
Can Step-3 replace Step-5-Pre?
It depends on your primary workload. If it sits in Step-3’s strengths (Edge-cloud products, In-car multimodal interaction, Light speech-plus-vision apps), yes. If you need A flagship-class stand-in for edge-cloud products, switching will cost quality.
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.