The new cheap tiers: $0.20/$1.20 versus $0.30/$2.50.
gpt-5.6-luna · openai
gemini-3.5-flash-lite · google
Benchmark
| Workload Scenario | GPT-5.6 Luna | Gemini 3.5 Flash-Lite | Price Delta |
|---|---|---|---|
| 1M input tokens (raw text) | $0.20 | $0.30 | −$0.10 |
| 1M output tokens (generation) | $1.20 | $2.50 | −$1.30 |
| 1M tokens · 70% input / 30% output mix | $0.50 | $0.96 | −$0.46 |
| Standard chat turn (4K in / 800 out, 50% cached) | $0.0014 | $0.00266 | −$0.00126 |
| Monthly scale (10K requests / day) | $420.00 | $798.00 | −$378.00 |
Negative difference = GPT-5.6 Luna is cheaper. Positive = Gemini 3.5 Flash-Lite is cheaper.
Scaling Curve
Total cost of a token volume at 70/30 input/output split (uncached). Log-log scale.
Verdict
GPT-5.6 Luna is cheaper on both input and output rates, so it costs less at every input/output mix. Price alone still isn't the whole decision: capability, latency and context limits (GPT-5.6 Luna: 1.1M, Gemini 3.5 Flash-Lite: 1M) may justify the premium for your task.
FAQ
On input, GPT-5.6 Luna is cheaper ($0.20/M vs $0.30/M). On output, GPT-5.6 Luna is cheaper ($1.20/M vs $2.50/M). The same model is cheaper on both sides, so it wins at every mix.
A chat-style request (4,000 input + 800 output tokens, 50% cached) costs $0.0014 on GPT-5.6 Luna and $0.00266 on Gemini 3.5 Flash-Lite — GPT-5.6 Luna is 1.9× cheaper for that workload.
GPT-5.6 Luna supports 1,050,000 tokens (128K max output); Gemini 3.5 Flash-Lite supports 1,048,576 (65.5K max output). GPT-5.6 Luna fits 1.0× more context, which matters for long documents and agents.