gemini-2.5-pro vs gemini-2.5-flash
gemini-2.5-pro vs gemini-2.5-flash: $1.25/$10 vs $0.3/$2.5 per 1M, the same workload priced on both; gemini-2.5-flash is 75% cheaper per request.
| gemini-2.5-pro | gemini-2.5-flash | |
|---|---|---|
| Provider · tier | gemini · frontier | gemini · fast |
| Input | $1.25 | $0.3 |
| Cached input | $0.125 | $0.03 |
| Cache write | none | none |
| Output | $10 | $2.5 |
| Long-context tier | > 200,000: $2.5 / $15 | none |
| Per request (2,000 in / 300 out) | $0.0055 | $0.00135 |
| Per month at 100k requests | $550 | $135 |
| …with the static prefix cached | $381 | $95 |
| Cheaper sibling (cascade) | gemini-2.5-flash | gemini-2.5-flash-lite |
Through TokenZero
Point your SDK at the gateway and keep both keys; the cache formatter, response cache and advisory routing act on every request, and each one comes back with a signed receipt showing the list-price baseline, actual cost and the saving.
OPENAI_BASE_URL=https://tokenzero.sh/w/<key>/v1 ANTHROPIC_BASE_URL=https://tokenzero.sh/w/<key>
Questions
Which is cheaper, gemini-2.5-pro or gemini-2.5-flash?
At list prices (table 2026-10-26) gemini-2.5-flash is 75% cheaper per request on a 2,000-in / 300-out workload. Quality is a separate question: TokenZero's cascade only routes to a sibling it scores as sufficient and measures parity with shadow evaluation.
Where do these prices come from?
The providers' own pricing pages, re-verified on 2026-10-26; the gateway signs receipts with the same table.