Qwen Plus 0728 (thinking) vs DeepSeek V4 Pro
A side-by-side comparison of two llm models — real specs, pricing, strengths and weaknesses, and a clear verdict on which to choose. Kept current by our agents.
Quick verdict: which should you choose?
Choose Qwen Plus 0728 (thinking) if you need
- ✓Choose DeepSeek V4 Pro if you need a reported intelligence_index of 51.5 and output speed of 79.81 t/s.
- ✓Choose DeepSeek V4 Pro if you need the largest context window at 1,048,576 tokens.
- ✓Choose DeepSeek V4 Pro if you need strong coding performance and structured outputs in technical domains.
- ✓Choose DeepSeek V4 Pro if you need explicit STEM-domain strengths alongside long-input handling.
Choose DeepSeek V4 Pro if you need
- ✓Choose Qwen Plus 0728 (thinking) if you need the lower output price of $0.78 per million tokens.
- ✓Choose Qwen Plus 0728 (thinking) if you need strong Chinese-English bilingual performance.
- ✓Choose Qwen Plus 0728 (thinking) if you need clear step-by-step reasoning style.
- ✓Choose Qwen Plus 0728 (thinking) if you need solid technical and coding assistance with long inputs.
Verdict
DeepSeek V4 Pro leads on measurable performance with its reported intelligence_index of 51.5 and output speed of 79.81 t/s, plus a marginally larger context window, while Qwen Plus 0728 (thinking) undercuts it on price at $0.78 per million tokens versus $0.87. Both models are open-weight, text-only, and excel at long-context technical work, but DeepSeek shows clearer edges in coding and STEM while Qwen emphasizes bilingual and step-by-step reasoning strengths. The data favor DeepSeek where concrete metrics exist and Qwen where cost is the priority.
Qwen Plus 0728 (thinking) vs DeepSeek V4 Pro: side by side
| Spec | Qwen Plus 0728 (thinking) | DeepSeek V4 Pro | Winner |
|---|---|---|---|
| Intelligence | — | 51.5 | Tie |
| Output speed | — | 80 t/s | Tie |
| Output price | $0.78/1M | $0.87/1M | Qwen Plus 0728 (thinking) |
| Context | 1000K | 1049K | DeepSeek V4 Pro |
| Params | — | — | Tie |
| Type | Open-weight | Open-weight | Tie |
| Provider | Alibaba Qwen | DeepSeek | Tie |
Detailed analysis
Performance
Winner: DeepSeek V4 ProDeepSeek V4 Pro provides a concrete intelligence_index of 51.5 and lists strong coding and STEM performance. Qwen Plus 0728 (thinking) lists solid technical and coding assistance plus step-by-step reasoning but supplies no intelligence_index. The single numeric benchmark therefore favors DeepSeek V4 Pro.
Pricing
Winner: Qwen Plus 0728 (thinking)Qwen Plus 0728 (thinking) lists output_price at $0.78 per million tokens. DeepSeek V4 Pro lists $0.87 per million tokens. The lower listed price gives Qwen Plus 0728 (thinking) the edge on cost.
Speed & Context
Winner: DeepSeek V4 ProDeepSeek V4 Pro reports output_speed of 79.81 t/s and context of 1,048,576 tokens. Qwen Plus 0728 (thinking) supplies neither speed nor an exact context figure beyond one million tokens. The available numeric data therefore favor DeepSeek V4 Pro on both speed and context size.
Domain Strengths
Winner: TieBoth models list effective long-input handling and technical or coding capabilities. DeepSeek V4 Pro adds emphasis on coding tasks and STEM domains while Qwen Plus 0728 (thinking) adds bilingual performance and step-by-step reasoning. The listed strengths are complementary rather than directly comparable.
Qwen Plus 0728 (thinking)
Pros
- +Strong Chinese-English bilingual performance
- +Effective handling of very long inputs
- +Solid technical and coding assistance
- +Clear step-by-step reasoning style
Cons
- –Text-only modality
- –May still hallucinate on niche facts
- –Performance varies across domains
DeepSeek V4 Pro
Pros
- +Strong performance on coding tasks
- +Effective handling of very long inputs
- +Clear and structured outputs
- +Good at technical and STEM domains
Cons
- –Text-only modality
- –No real-time information access
- –Can produce hallucinations on facts
Summary: Qwen Plus 0728 (thinking) vs DeepSeek V4 Pro
Select DeepSeek V4 Pro when the published intelligence_index, speed, or maximum context length matters most. Select Qwen Plus 0728 (thinking) when the lower listed price or bilingual reasoning style is the deciding factor. Both remain viable open-weight options for long-context technical work.
Frequently asked questions
DeepSeek V4 Pro is better where concrete metrics are available because it alone supplies an intelligence_index of 51.5 and output speed of 79.81 t/s.