Head to head
claude-haiku-4-5 vs gemini-3.5-flash-lite: API Pricing Compared
On a typical chat turn of 10K input and 1K output tokens, gemini-3.5-flash-lite costs $0.00550 against $0.015 for claude-haiku-4-5, making it 63% cheaper. Both rates are read daily from the providers' own pricing pages.
Side by side
| claude-haiku-4-5 | gemini-3.5-flash-lite | |
|---|---|---|
| Provider | Anthropic | |
| Input / 1M tokens | $1.00 | $0.30 |
| Output / 1M tokens | $5.00 | $2.50 |
| Context window | 200K | Not published |
| Blended cost / chat turn | $0.015 | $0.00550 |
| Price verified | 2026-06-11 | 2026-08-06 |
Cost by workload
The headline rate rarely decides the bill. These are the same four workloads priced against both models.
| Workload | Tokens | claude-haiku-4-5 | gemini-3.5-flash-lite |
|---|---|---|---|
| Single short call | 2K in, 500 out | $0.00450 | $0.00185 |
| Typical chat turn | 10K in, 1K out | $0.015 | $0.00550 |
| Long document pass | 100K in, 2K out | $0.110 | $0.035 |
| 1,000 chat turns | 10M in, 1M out | $15.00 | $5.50 |
Which one actually wins
gemini-3.5-flash-lite is cheaper on both the input and the output rate, so it never costs more no matter how input-heavy or output-heavy the workload is. There is no crossover point to reason about. Price is only one axis, so weigh it against context window, latency, and the quality difference on your own evaluation set before switching.
Common questions
Which is cheaper, claude-haiku-4-5 or gemini-3.5-flash-lite?
gemini-3.5-flash-lite is cheaper for a typical chat turn of 10K input and 1K output tokens, at $0.00550 against $0.015 for claude-haiku-4-5, a difference of 63%.
How much does claude-haiku-4-5 cost per 1M tokens?
Anthropic charges $1.00 per 1M input tokens and $5.00 per 1M output tokens for claude-haiku-4-5, verified 2026-06-11.
How much does gemini-3.5-flash-lite cost per 1M tokens?
Google charges $0.30 per 1M input tokens and $2.50 per 1M output tokens for gemini-3.5-flash-lite, verified 2026-08-06.
Does gemini-3.5-flash-lite stay cheaper for every workload?
Yes. gemini-3.5-flash-lite is cheaper on both the input and the output rate, so it never costs more regardless of how input-heavy or output-heavy the workload is.
Full pricing pages
Maintained by Jordan Hale, TechFastForward. Both prices are read daily from the official provider pages and recorded only when they move. Compare any other pair on the LLM API Pricing Tracker.