gpt-5 vs llama-4-maverick
On a 1,500-input, 500-output call at 10,000 calls a day, gpt-5 costs about $3,150 a month and llama-4-maverick about $300 — llama-4-maverick is roughly 90% cheaper than gpt-5. gpt-5 lists $12.00 per 1M output tokens against llama-4-maverick at $0.50.
gpt-5 costs $3.00/1M in and $12.00/1M out; llama-4-maverick costs $0.50/1M in and $0.50/1M out.
Last updated . Model pricing is refreshed twice daily.
Price per 1M tokens
gpt-5: input $3.00, output $12.00. llama-4-maverick: input $0.50, output $0.50.
- Monthly at 10,000 calls/day — gpt-5: $3,150, llama-4-maverick: $300
- Monthly at 100,000 calls/day — gpt-5: $31,500, llama-4-maverick: $3,000
- Time to first token — gpt-5: 250 ms, llama-4-maverick: 280 ms
- Throughput — gpt-5: 180 tok/s, llama-4-maverick: 55 tok/s
Which one to run
If your workload tolerates llama-4-maverick's quality, the saving is immediate. If it does not, route: send the easy calls to llama-4-maverick and keep gpt-5 for the hard ones.
Frequently asked questions
Is gpt-5 cheaper than llama-4-maverick?
llama-4-maverick is cheaper. At 10,000 calls a day with 1,500 input and 500 output tokens, gpt-5 costs about $3,150 a month and llama-4-maverick about $300.
What is the price difference between gpt-5 and llama-4-maverick?
About 90% on the same workload, using list prices with no caching or batch discounts.