Gemini 2.0 Models by Google DeepMind
Google DeepMindHighlight
9 models2024–2025Up to 2m ctxFrom $0.075/1M input
About
Gemini 2.0 is a family of 9 AI models by Google DeepMind, released between 2024 and 2025.
Current Variants
Use-when guidance is derived from seed capabilities, context, release, and replacement fields.
5 in view4 retired
Use when the workload needs robotics, 32k context, and tool use.
2025-03robotics32k contexttool use
Gemini 2.0 Flash Live APICurrent
Use when the workload needs 1m context, tool use, and function calling.
2025-031m contexttool usefunction calling
Use when the workload needs 1m context.
2025-021m context
Use when the workload needs 1m context and structured outputs.
2025-021m contextstructured outputs
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Gemini Robotics-ER 1.5 Preview | Use when the workload needs robotics, 32k context, and tool use. | 2025-03 | robotics32k contexttool use | Current |
| Gemini 2.0 Flash Live API | Use when the workload needs 1m context, tool use, and function calling. | 2025-03 | 1m contexttool usefunction calling | Current |
| Gemini 2.0 Flash-Lite (Preview 02-05) | Use when the workload needs 1m context. | 2025-02 | 1m context | Current |
| Gemini 2.0 Pro (Experimental 02-05) | Use when the workload needs 1m context and structured outputs. | 2025-02 | 1m contextstructured outputs | Current |
| Gemini 2.0 Flash Experimental | Use when the workload needs 1m context. | 2024-12 | 1m context | Current |
Release Timeline
4 release groups2025-03
2 current
Gemini 2.0 Flash Live API
Current1m contexttool usefunction calling
Gemini Robotics-ER 1.5 Preview
Currentrobotics32k contexttool use
2025-02
2 current · 3 retired
Gemini 2.0 Flash Image Generation
Archivedimage1.05m contexttool use
Gemini 2.0 Flash Lite
Replaced1m contextstructured outputs
Gemini 2.0 Flash-Lite
Replaced1.05m contexttool usefunction calling
Gemini 2.0 Flash-Lite (Preview 02-05)
Current1m context
Gemini 2.0 Pro (Experimental 02-05)
Current1m contextstructured outputs
2025-01
1 retired
Gemini 2.0 Flash
Replaced2m contextreasoningtool use
2024-12
1 current
Gemini 2.0 Flash Experimental
Current1m context
Replaced By
Keep for legacy integrations; evaluate Gemini 3.1 Flash-Lite before new work.
Keep for legacy integrations; evaluate Gemini 3.1 Flash-Lite before new work.
Replaced
Keep for legacy integrations; evaluate Gemini 3.5 Flash before new work.
Specifications(9 models)
| Model | Released | Context | Vision | Multimodal | Reasoning | Fn Calling | Tool Use | Structured Outputs | Code Exec |
|---|---|---|---|---|---|---|---|---|---|
| Gemini Robotics-ER 1.5 Preview | 2025-03 | 32k | Yes | Yes | No | Yes | Yes | No | No |
| Gemini 2.0 Flash Live API | 2025-03 | 1m | Yes | Yes | No | Yes | Yes | Yes | No |
| Gemini 2.0 Flash-Lite (Preview 02-05) | 2025-02 | 1m | No | No | No | No | No | No | No |
| Gemini 2.0 Pro (Experimental 02-05) | 2025-02 | 1m | No | No | No | No | No | Yes | No |
| Gemini 2.0 Flash Experimental | 2024-12 | 1m | No | No | No | No | No | No | No |
Available From(4 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| Gemini 2.0 Flash Live API | GCP Vertex AI | $0.5 | — | Serverless |
Frequently Asked Questions
- What is Gemini 2.0 used for?
- Gemini 2.0 is used for robotics, image, and vision and multimodal work. The family description and listed model capabilities point to those workloads as the best fit.
- How does Gemini 2.0 compare to Gemma 4?
- Gemini 2.0 by Google DeepMind is strongest where you need robotics, while Gemma 4 by Google DeepMind is the closest related family to check for multimodal. Gemini 2.0 has 9 listed variants and reaches up to 2m context, while Gemma 4 reaches up to 256k context, so compare the specs and pricing tables before choosing a production model.
- Which Gemini 2.0 model should I use?
- For the lowest listed input price, start with Gemini 2.0 Flash-Lite through Google AI Studio at $0.075/1M input tokens. For the most capable/latest local choice, evaluate Gemini 2.0 Flash Live API with 1m context and tool use, function calling, structured outputs, and multimodal inputs.
