Gemini 1.5 Pro-002 and Flash-002
Updated 1.5 models gain about +7% on MMLU-Pro and +20% on math, while 1.5 Pro input price is cut 64% and output price cut 52%.
Production-ready refresh with 2x faster output and 3x lower latency, higher rate limits (Flash 2,000 RPM, Pro 1,000 RPM) and looser default safety filters. Price cuts applied to prompts under 128K tokens from 2024-10-01.
- Date
- Tuesday, 24 September 2024
- Lab
- Google DeepMind
- Kind
- model
- Access
- closed API
- Price
- Price cut effective 2024-10-01
Figures
| Measure | Value | Measured by |
|---|---|---|
| 1.5 Pro price cut (prompts under 128K) | -64% input, -52% output effective 2024-10-01 | company |
Sources
This record was checked against its sources on 6 October 2026. How we check