Curiosity · AI Model
Gemini 1.5 Pro
Gemini 1.5 Pro is Google's February 2024 flagship and the model that made million-token context windows real. It introduced a mixture-of-experts architecture with 1M tokens standard and 2M in private preview, and was the first model to demonstrate reliable native video understanding across hour-long clips.
Model specs
- Vendor
- Family
- Gemini 1.5
- Released
- 2024-02
- Context window
- 2,000,000 tokens
- Modalities
- text, vision, audio, video, code
- Input price
- $1.25/M tok
- Output price
- $5/M tok
- Pricing as of
- 2026-04-20
Strengths
- First truly usable million-token context — near-perfect haystack recall
- Native video understanding, not just frame sampling
- MoE architecture delivers strong quality at Pro-tier cost
- Broad availability across Vertex AI, AI Studio, and Google Cloud
Limitations
- Superseded on benchmarks by Gemini 2.x and 2.5 Pro
- No reasoning/thinking mode
- Latency on full 1M context is slow — plan caching around it
Use cases
- Long-document and long-video analysis (hours of footage)
- Codebase-wide reasoning with a repo in context
- Lecture and meeting summarisation from raw audio/video
- Research tasks requiring many sources in one prompt
Benchmarks
| Benchmark | Score | As of |
|---|---|---|
| MMLU | ≈85.9% | 2024-05 |
| Video-MME | ≈75% | 2024-05 |
| Needle-in-a-haystack (1M) | ≈99% | 2024-02 |
Frequently asked questions
What is Gemini 1.5 Pro?
Gemini 1.5 Pro is Google's February 2024 flagship long-context model. It was the first widely-available model with a 1M-token context window (with a 2M preview) and popularised native video and audio understanding inside a single LLM.
What is the context window of Gemini 1.5 Pro?
Gemini 1.5 Pro supports a 1M-token context window in general availability, with a 2M-token window made available to select customers. It remains one of the largest context windows on any production model.
Is Gemini 1.5 Pro still worth using in 2026?
For new projects, Gemini 2.5 Pro or 2.5 Flash is usually a better choice — they are faster, smarter, and include thinking mode. Gemini 1.5 Pro remains useful for pipelines tuned to its behaviour and for workloads that rely on its 2M context offer.
How much does Gemini 1.5 Pro cost?
As of April 2026, Gemini 1.5 Pro is priced at roughly USD 1.25 per million input tokens and USD 5 per million output tokens on Vertex AI. Long-context pricing tiers may apply above 128K tokens.
Sources
- Google — Gemini 1.5 Pro — accessed 2026-04-20
- Google Cloud — Gemini pricing — accessed 2026-04-20