Cost by workload size
Price comparison
| Feature |  Claude Opus 5 |  GPT-4o |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Input / M tokens | $5 | $2.50 |
| Output / M tokens | $25 | $10 |
| Cached input read | $0.50 | $1.25 |
| Context window | 1M | 128K |
| Batch discount | — | −50% |
| Vision | Yes | Yes |
| Audio | — | Yes |
| Reasoning tokens | Yes | — |
What your workload actually costs
Prices per token hide the real picture. Here are three common workloads computed with AITokenCalculator's own engine (medium reasoning effort, no batch, no cache):
| Workload | Claude Opus 5 | GPT-4o | Cheaper |
|---|---|---|---|
| 1,000 in / 300 out | $0.02 | $0.0055 | GPT-4o |
| 10,000 in / 2,000 out | $0.15 | $0.045 | GPT-4o |
| 100,000 in / 20,000 out | $1.50 | $0.45 | GPT-4o |
| Monthly @ 1,000 req/day (standard) | $4,565.63 | $1,369.69 | GPT-4o |
Analysis
On input price, GPT-4o wins at $2.50/M — roughly 2.0× cheaper than its rival. Output tells a sharper story: GPT-4o charges $10/M , about 2.5× less than Claude Opus 5 — and output is where chatty and agentic workloads bleed money.
Context capacity differs too: Claude Opus 5 fits 1M tokens (~750,000 words) versus 128K. If you feed whole documents or codebases, that gap decides feasibility before price even matters.
Premium intelligence vs balanced cost. Opus-class models are the strongest reasoners Anthropic ships and price accordingly — expect several times GPT-4o’s output bill. Reserve it for agentic workflows and hard analysis where failure cost exceeds token cost.
Numbers shift as providers reprice — check the full AI model pricing table, then run your own prompt through the calculator preloaded with Claude Opus 5 or with GPT-4o.