Model comparison
DeepSeek-V4-Pro-0813 vs GPT-5.4 mini
A spec-by-spec comparison of DeepSeek-V4-Pro-0813 and GPT-5.4 mini: pricing, context window, capabilities, and independently sourced benchmarks.
Head-to-head
- DeepSeek-V4-Pro-0813 costs 1.1x less per input token than GPT-5.4 mini ($0.66/MTok vs $0.75/MTok).
- DeepSeek-V4-Pro-0813 has a 2.5x larger context window than GPT-5.4 mini (1,000,000 tokens vs 400,000 tokens).
- Only GPT-5.4 mini supports vision.
| Spec | DeepSeek-V4-Pro-0813 | GPT-5.4 mini |
|---|---|---|
| Provider | deepseek | openai |
| Family | DeepSeek V4 | GPT-5.4 |
| Status | frontier | current |
| Release date | — | — |
| Knowledge cutoff | — | 2025-08-31 |
| Context window | 1,000,000 tokens | 400,000 tokens |
| Max output tokens | 384,000 tokens | 128,000 tokens |
| Input price | $0.66/MTok | $0.75/MTok |
| Output price | $1.98/MTok | $4.50/MTok |
| Cached input price | $0.02/MTok | $0.07/MTok |
| Vision | No | Yes |
| Tool use | Yes | Yes |
| Extended thinking | Yes | Yes |
| Streaming | — | Yes |
DeepSeek-V4-Pro-0813
Strengths- Cache-hit input at $0.022/MTok (off-peak) is roughly 30x cheaper than this model’s own off-peak standard input rate.
- 384K max output is triple the 128K ceiling of the OpenAI and Anthropic flagships.
- Thinking and non-thinking modes are billed at one price.
- Exposes both OpenAI-format and Anthropic-format endpoints.
- Peak-hour pricing (01:00-04:00 and 06:00-10:00 UTC, Monday-Friday, excluding Chinese public holidays) doubles every rate: $1.32/MTok input, $3.96/MTok output, $0.044/MTok cached.
- Vendor label is now DeepSeek-V4-Pro-0813, a dated snapshot; the vendor page previously carried an undated label.
- Does not support image input.
- Concurrency is capped at 500 requests.
- DeepSeek states it reserves the right to adjust prices and recommends checking the page regularly.
GPT-5.4 mini
Strengths- Described by OpenAI as its strongest mini model for coding, computer use and subagents.
- 400K context window, confirmed on its own dedicated model page.
- The rate card publishes no long-context row, so behaviour above the short-context threshold is undocumented.
Not sure which one to use?
Flaibo is a daily AI orientation hub: a 12-week path and a prompt vault, with Free and Professional plans. Build the habit, not just the comparison.
Start your AI path