Google Gemini
Fine-tune Gemini 2.5 Flash, Flash Lite, and Pro on Commissioned.
Gemini models excel at long-context tasks and document analysis. Fine-tuning runs on Google Cloud infrastructure and takes 30–45 minutes.
Available models
Google's general-purpose model with excellent long-context handling. Handles very large documents well.
Strengths:
- Large context window — good for long documents
- Strong at summarization and extraction
- Competitive quality with GPT-4.1 Mini
Good for: research, document analysis, long-form content, knowledge bases
Plan: Free
The lightest Gemini model. Optimized for speed and simplicity.
Strengths:
- Fastest inference in the Gemini family
- Good enough for straightforward tasks
- Lowest resource usage
Good for: high-throughput applications, simple classification, quick Q&A
Plan: Free
Google's most capable model. The best Gemini option when quality is the priority.
Strengths:
- Best reasoning in the Gemini family
- Excellent at complex, multi-step tasks
- Strong creative and analytical writing
Good for: premium assistants, production use cases, tasks where quality matters more than cost
Plan: Pro ($25/mo)
Gemini 2.5 Pro requires a Pro plan. All other Gemini models are available on the free tier.
Training details
| Flash | Flash Lite | Pro | |
|---|---|---|---|
| Training time | 30–45 min | 30–45 min | 30–45 min |
| Output | Hosted by Google | Hosted by Google | Hosted by Google |
| Self-hosting | Not available | Not available | Not available |
| Plan required | Free | Free | Pro |