DeepSeek V4 Pro vs GPT-6 Astra: price
DeepSeek V4 Pro is 10.1× cheaper than GPT-6 Astra for a typical chat workload (1,500 tokens in, 500 out).
| Item | DeepSeek V4 Pro | GPT-6 Astra |
|---|---|---|
| Input per 1M tokens | $1.32 | $10.00 |
| Output per 1M tokens | $3.96 | $50.00 |
| Cached input per 1M | $0.044 | $1.00 |
| 1,000 chat messages1,500 tokens in, 500 out each | $3.96 | $40.00 |
| Summarize a 50-page PDF30,000 tokens in, 1,000 out | $0.044 | $0.35 |
| One coding-agent task60,000 tokens in, 4,000 out, no caching | $0.095 | $0.80 |
DeepSeek V4 Pro: Peak-hour price (weekdays 01:00–04:00 and 06:00–10:00 UTC). Off-peak: $0.66 in / $1.98 out.
GPT-6 Astra: Prompts over 272K tokens: 2x input, 1.5x output.
List prices checked September 28, 2026. Price is half the story: see how people rate each one today.
DeepSeek V4 Pro vs GPT-6 Astra in numbers
Output, usually the bigger part of the bill, costs $3.96 on DeepSeek V4 Pro and $50.00 on GPT-6 Astra, so DeepSeek V4 Pro is 12.6× cheaper there. Input costs $1.32 versus $10.00 per million tokens. For a coding-agent task (60,000 tokens in, 4,000 out) that is $0.095 on DeepSeek V4 Pro and $0.80 on GPT-6 Astra. DeepSeek V4 Pro accepts 1M tokens of context and GPT-6 Astra 1.05M tokens. DeepSeek V4 Pro came out on August 13, 2026 and GPT-6 Astra on September 3, 2026.
DeepSeek V4 Pro is made by DeepSeek; GPT-6 Astra by OpenAI. Estimate your own bill for both in the cost calculator.
Which one feels dumber today?
- DeepSeek V4 ProDeepSeek–no votes yet
- GPT-6 AstraOpenAI–no votes yet
Frequently asked questions
Is DeepSeek V4 Pro or GPT-6 Astra cheaper?
DeepSeek V4 Pro is 10.1× cheaper than GPT-6 Astra for a typical chat workload (1,500 tokens in, 500 out). Output, usually the bigger part of the bill, costs $3.96 on DeepSeek V4 Pro and $50.00 on GPT-6 Astra, so DeepSeek V4 Pro is 12.6× cheaper there.
Which is better, DeepSeek V4 Pro or GPT-6 Astra?
It depends on the task, and price is only part of it. Check how people rate each one today on the DeepSeek V4 Pro and GPT-6 Astra pages, and read what users report in Discussions.