Claude Haiku 4.5 vs Gemini 2.5 Pro
A direct API-pricing comparison between Claude Haiku 4.5 and Gemini 2.5 Pro. Below: every cost dimension side by side, then the bottom line for a typical workload.
Bottom line
For a typical 5K-in / 1K-out request, Claude Haiku 4.5 is the cheaper choice at $0.010 per call — 1.6× less than Gemini 2.5 Pro at $0.016.
| Claude Haiku 4.5Anthropic | Gemini 2.5 ProGoogle | |
|---|---|---|
| Input / 1M tokens | $1.00 | $1.25 |
| Output / 1M tokens | $5.00 | $10.00 |
| Cached input / 1M | $0.100 | $0.310 |
| Context window | 200K | 1M |
| Provider | Anthropic |
Green = cheaper on that line. Prices are estimates from public pricing pages; verify against Anthropic and Google before relying on them.
Which is cheaper, Claude Haiku 4.5 or Gemini 2.5 Pro?
On output tokens — usually the dominant cost — Claude Haiku 4.5 charges $5.00 per million versus $10.00 for Gemini 2.5 Pro. The right pick depends on your input/output ratio and how much you can serve from cache. Use the calculator to model your exact traffic.