DeepSeek V4 Pro 0813 vs GPT-6 Luna: cost per task
GPT-6 Luna is cheaper on 14 of 14 tasks and about 14.5x cheaper on average (geometric mean).
| Task | DeepSeek V4 Pro 0813 | GPT-6 Luna | Difference |
|---|---|---|---|
| Answer a chat question | 1.3¢ | 0.06¢ | 21.1x |
| Reply to a customer email | 0.41¢ | <0.01¢ | 56.2x |
| Write a 1,500-word blog post | 0.61¢ | 0.12¢ | 5.1x |
| Write 10 product descriptions | 0.56¢ | 0.05¢ | 10.3x |
| Translate 1,000 words | 2.4¢ | 0.08¢ | 29.0x |
| Proofread one page (500 words) | 0.57¢ | 0.04¢ | 12.6x |
| Summarize a 20-page PDF | 1.3¢ | 0.16¢ | 8.0x |
| Extract data from an invoice | 0.60¢ | 0.03¢ | 18.8x |
| Analyze a 200-row spreadsheet | 3.5¢ | 0.19¢ | 18.6x |
| Answer from 5 retrieved documents (RAG) | 0.99¢ | 0.05¢ | 18.3x |
| Classify 100 support tickets | 1.1¢ | 0.06¢ | 16.5x |
| Review 500 lines of code | 3.5¢ | 0.31¢ | 11.3x |
| Write a SQL query | 0.43¢ | 0.04¢ | 10.7x |
| Run a 50-step coding agent | 30¢ | 4.1¢ | 7.5x |
Specs and list prices
| DeepSeek V4 Pro 0813 | GPT-6 Luna | |
|---|---|---|
| Input per 1M | $0.66 | $0.10 |
| Output per 1M | $1.98 | $0.50 |
| Cached input per 1M | $0.022 | $0.01 |
| Context | 1M | 1.05M |
| Providers | 23 | 7 |
| Reasoning | default | default |
| LMArena Text | - | 1443 |
| Released | Aug 12, 2026 | Sep 22, 2026 |
FAQ
Is DeepSeek V4 Pro 0813 or GPT-6 Luna cheaper?
GPT-6 Luna is cheaper on 14 of 14 tasks and about 14.5x cheaper on average (geometric mean).
How much does DeepSeek V4 Pro 0813 cost compared to GPT-6 Luna per token?
DeepSeek V4 Pro 0813: $0.66 input and $1.98 output per 1M tokens. GPT-6 Luna: $0.10 input and $0.50 output. Token prices alone mislead - models use different numbers of tokens for the same task, which is what the task table shows.