gemini-2.5-flash-preview-09-2025| Provider | Input, $ per 1M | Output, $ per 1M | In our data since |
|---|---|---|---|
| 302.AI | $0.3 | $2.5 | 31 Jul 2026 |
| Google DeepMind | $0.3 | $2.5 | 31 Jul 2026 |
| Google Vertex AI | $0.3 | $2.5 | 31 Jul 2026 |
| NanoGPT | $0.3 | $2.5 | 31 Jul 2026 |
Only the base tier and only per-token prices are shown. Batch, discounted and cached rates, as well as prices per image or per second of video, do not go into this table: they cannot stand in the same column as a price per million tokens.
The date in the last column is the day this price first entered our collection. The price may well be older: before that day we simply were not recording it. It has not changed since — otherwise a new row with a new date would stand in its place.
| Task set | Result | Run conditions | Measured by |
|---|---|---|---|
| Arena Score in French | 1,437.16 1,411–1,464 | — | — |
| Arena Score, expert questions | 1,425.27 1,411–1,440 | — | — |
| Arena Score, mathematics | 1,416.44 1,403–1,430 | — | — |
| Arena Score in English | 1,411.08 1,406–1,416 | — | — |
| Arena Score, hard prompts | 1,409.13 1,404–1,414 | — | — |
| Arena Score, overall | 1,406.92 1,403–1,411 | — | — |
| Arena Score, long queries | 1,405.75 1,399–1,413 | — | — |
| Arena Score, programming | 1,401.55 1,394–1,409 | — | — |
| Arena Score in Spanish | 1,398.72 1,377–1,420 | — | — |
| Arena Score in Russian | 1,398.71 1,387–1,411 | — | — |
| Arena Score, multi-turn dialogue | 1,395.45 1,387–1,403 | — | — |
| Arena Score, instruction following | 1,394.54 1,388–1,401 | — | — |
| Arena Score, creative writing | 1,387.99 1,379–1,397 | — | — |
| Arena Score, text recognition in images | 1,255.92 1,244–1,268 | — | — |
| Arena Score, understanding diagrams | 1,249.38 1,231–1,268 | — | — |
| Arena Score, working with images | 1,248.77 1,239–1,259 | — | — |
| WeirdML — unusual machine learning tasks | 41.91 % | thinking budget: 16K | WeirdML Leaderboard |
| The Agent Company — work tasks in an office environment | 41.1 % | — | TheAgentCompany leaderboard |
| Terminal-Bench — working in the command line | 17.1 % | · with a tuned harness | Terminal-Bench v2 Leaderboard |