GPT-4o vs GPT-6 Luna: cost

GPT-4o: ✓ verified 2026-11-19 · Official pricing ↗
GPT-6 Luna: ✓ verified 2026-11-19 · Official pricing ↗
Synced 2026-11-19 01:00 UTC · Method

For a typical support chatbot, GPT-6 Luna is about 23.8× cheaper than GPT-4o retiring 2027-01-14 ($38.51 vs $915 per month). Price is only one factor: test both on your own prompts before switching.

Monthly cost by workload

Estimated monthly cost by workload (USD)
Scenario (per month)GPT-4oGPT-6 Luna
Customer support chatbot
5,000 req/day · 1,200 in · 400 out · cache 60%
$915$38.51
RAG over documents
2,000 req/day · 8,000 in · 600 out · cache 30%
$1,380$53.64
Coding agent
1,500 req/day · 25,000 in · 2,000 out · cache 80%
$2,588$79.31
Deep reasoning
800 req/day · 3,000 in · 6,000 out · cache 20%
$1,602$77.90
Batch processing
50,000 req/day · 1,500 in · 150 out · batch
$3,938$169
GPT-4o vs GPT-6 Luna: prices and limits
FieldGPT-4o retiring 2027-01-14GPT-6 Luna
Input /1M$2.50$0.10
Output /1M$10.00$0.50
Cached /1M$1.25$0.01
Batch input /1M$1.25$0.05
Batch output /1M$5.00$0.25
Context window128,0001,050,000
Max input128,000922,000
Max output16,384128,000

Where the prices differ

Which one to pick

Pick GPT-6 Luna when cost per request dominates: high-volume chatbots, classification, extraction, or anything you can run through a Batch API. GPT-4o retiring 2027-01-14 costs more without a larger context window or extra capabilities in our data; choose it only if it scores better on your own evaluation.

Add a third model or change the workload in the compare tool, or open the model pages: GPT-4o retiring 2027-01-14, GPT-6 Luna.

FAQ

Which is cheaper, GPT-4o or GPT-6 Luna?

GPT-6 Luna. For the support-chatbot scenario it costs $38.51 per month against $915 for GPT-4o retiring 2027-01-14, about 23.76× less; on input tokens the gap is 25× and on output 20×.

What do 1M output tokens cost on GPT-4o and GPT-6 Luna?

GPT-4o retiring 2027-01-14: $10.00 per 1M output tokens ($2.50 input). GPT-6 Luna: $0.50 per 1M output ($0.10 input). Prices are checked every 4 hours; see each model page for cache, batch and long-context rates.

Which has the larger context window?

GPT-6 Luna, with 1,050,000 tokens against 128,000 for GPT-4o retiring 2027-01-14. Prompts near the limit cost more in absolute terms, so check the long-context rates on the model pages.