DeepSeek-V4-Pro-0813 vs GPT-6 Astra: Specs & Benchmark Comparison
DeepSeek-V4-Pro-0813 is developed by DeepSeek, while GPT-6 Astra comes from OpenAI. DeepSeek-V4-Pro-0813 was released in August 2026, and GPT-6 Astra followed a month later in September 2026. DeepSeek-V4-Pro-0813 has a published size of about 1.6 trillion parameters; OpenAI has not disclosed the parameter count of GPT-6 Astra.
The two models share 3 published benchmarks. GPT-6 Astra leads on 2 of them, DeepSeek-V4-Pro-0813 on 1. The widest gaps are on Agents' Last Exam, where GPT-6 Astra scores 59.3% against 25.7%; AutomationBench, where GPT-6 Astra scores 41.4% against 31.8%. Averaged across everything we track, DeepSeek-V4-Pro-0813 sits at 60.6% and GPT-6 Astra at 74.4%.
DeepSeek-V4-Pro-0813 is the cheaper API at $1.32 per million input tokens and $3.96 per million output tokens, roughly 8 times cheaper than GPT-6 Astra at $10 and $50. Both accept a context window of about 1M tokens. DeepSeek-V4-Pro-0813 accepts text as input, while GPT-6 Astra accepts text and images.
| Characteristic | DeepSeek-V4-Pro-0813 | GPT-6 Astra |
|---|---|---|
| Company | DeepSeek | OpenAI |
| Release Date | August 13, 2026 | September 4, 2026 |
| Parameters | 1.6T | — |
| Multimodal | No | Yes |
| Context (input) | 1.0M | 1.1M |
| Context (output) | 393K | 128K |
| Input Price / 1M | $1.32 | $10.00 |
| Output Price / 1M | $3.96 | $50.00 |
| Average Score | 60.6% | 74.4% |
| Benchmarks | ||
| Agents' Last Exam | 25.7% | 59.3% |
| AutomationBench | 31.8% | 41.4% |
| Humanity's Last Exam (with tools, text-only) | 60.0% | 57.2% |
Visual Benchmark Comparison
Verdict
GPT-6 Astra leads in 2 out of 4 comparison categories.
Both models show comparable average scores: DeepSeek-V4-Pro-0813 — 0.6, GPT-6 Astra — 0.7.
DeepSeek-V4-Pro-0813 is 11.4x cheaper: input $1.32/1M vs $10.00/1M tokens.
GPT-6 Astra supports a larger context: 1M vs 1M tokens.
GPT-6 Astra is newer: released 9/4/2026 vs 8/13/2026.
More About These Models
Related Comparisons
Frequently Asked Questions
Which is better for coding — DeepSeek-V4-Pro-0813 or GPT-6 Astra?
Which model is cheaper — DeepSeek-V4-Pro-0813 or GPT-6 Astra?
Which has a larger context window — DeepSeek-V4-Pro-0813 or GPT-6 Astra?
The DeepSeek-V4-Pro-0813 and GPT-6 Astra comparison is updated for 2026. Data includes benchmark results, API pricing, context window size and other specifications. For more detailed information, visit the DeepSeek-V4-Pro-0813 or GPT-6 Astra page. See also the complete list of AI model comparisons.