Claude Opus 4.6 vs GPT-5.4
Anthropic against OpenAI, costed on the same three workloads. GPT-5.4 is cheaper on all three workloads.
Monthly bill, side by side
| Workload | Claude Opus 4.6 | GPT-5.4 | Difference |
|---|---|---|---|
| Support chatbot 2,500 six-turn conversations a day |
$20,138 | $11,081 | $9,056 (1.8×) |
| Document summariser 1,000 single-shot summaries a day |
$2,430 | $1,275 | $1,155 (1.9×) |
| Coding agent 200 twelve-turn sessions a day |
$14,040 | $7,470 | $6,570 (1.9×) |
No prompt caching and no batch discount on either side, so the comparison is like for like. Both figures assume 30 days a month.
The rate cards
| Claude Opus 4.6 | GPT-5.4 | |
|---|---|---|
| Input, per 1M tokens | $5 | $2.5 |
| Output, per 1M tokens | $25 | $15 |
| Cached input | $0.5 | $0.25 |
| Context window | 1M | 1.05M |
| Maximum output | 128K | 128K |
| Reasoning model | yes | yes |
| Vision | yes | yes |
Run this comparison on your own numbers
Full detail: Claude Opus 4.6 pricing · GPT-5.4 pricing