gemini-2.5-flash-preview-04-17| Provider | Input, $ per 1M | Output, $ per 1M | In our data since |
|---|---|---|---|
| NanoGPT | $0.15 | $0.6 | 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 |
|---|---|---|---|
| Creative writing (Lech Mazur’s evaluation) | 76.5 % | thinking budget: 24K | lechmazur/writing Github repository |
| Mock AIME 2024–2025 — olympiad problems | 73.03 % | · with a tuned harness | Epoch evaluations |
| GeoBench — locating a place from a photograph | 73 % | — | GeoBench leaderboard |
| Fiction.LiveBench — holding a long context | 47.2 % | — | Fiction.live leaderboard |
| Aider Polyglot — code edits in six languages | 47.1 % | · with a tuned harness | Aider LLM Leaderboards |
| WeirdML — unusual machine learning tasks | 40.95 % | thinking budget: 16K | WeirdML Leaderboard |
| ARC-AGI — generalising to unseen patterns | 32.3 % | thinking budget: 24K · with a tuned harness | ARC Prize Leaderboard |
| DeepResearch Bench — deep research | 29.19 % | — | — |
| Humanity’s Last Exam — expert-level questions | 7.65 % | — | — |