GPT-6 Luna vs Gemini 3.1 Flash-Lite: pricing compared
For a 30-step coding agent run, GPT-6 Luna is 2.5x cheaper ($0.05 vs $0.13). Prices verified 2026-10-18.
| GPT-6 Luna | Gemini 3.1 Flash-Lite | |
|---|---|---|
| Provider | OpenAI | |
| Input / 1M | $0.10 | $0.25 |
| Cached input / 1M | $0.01 | $0.025 |
| Output / 1M | $0.50 | $1.50 |
| Batch discount | 50% | 50% |
| Context window | 1.05M | 1.05M |
| Long-context tier | above 272K | none |
Cost on real workloads
| Workload | GPT-6 Luna | Gemini 3.1 Flash-Lite |
|---|---|---|
| 10,000 chatbot replies 2,000 input + 400 output tokens each, 50% of the prompt cached | $3.10 | $8.75 |
| One 30-step coding agent run 12K system + tools prompt, 1K output and 3K tool result per step, caching on | $0.05 | $0.13 |
| 1M documents classified (batch) 600 input + 20 output tokens each, batch API where available | $35.00 | $90.00 |
Price is only half the decision: a cheaper model that needs more steps or retries can cost more per finished task. Run your own numbers in the calculator (both models pinned), or see how both rank in the use-case guides.
FAQ
Is GPT-6 Luna cheaper than Gemini 3.1 Flash-Lite?
For a typical 30-step coding agent run with caching, GPT-6 Luna costs $0.05 and Gemini 3.1 Flash-Lite costs $0.13, so GPT-6 Luna is 2.5x cheaper. Per 1M tokens, GPT-6 Luna is $0.10 / $0.50 and Gemini 3.1 Flash-Lite is $0.25 / $1.50.
Which has the bigger context window?
GPT-6 Luna: 1,050,000 tokens. Gemini 3.1 Flash-Lite: 1,048,576 tokens.