Gemma 4 Models by Google DeepMind
Details
Capabilities
About
Google's most capable open-source model family, purpose-built for advanced reasoning and agentic workflows. Delivered in five sizes (E2B, E4B, 12B dense, 26B MoE, 31B dense) with multimodal capabilities including text, image, video, and audio processing.
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
Use when the workload needs 256k context, 12B parameters, and reasoning.
Use when the workload needs 256k context, 12B parameters, and reasoning.
Use when the workload needs 128k context, 2B parameters, and function calling.
Use when the workload needs 128k context, 2B parameters, and function calling.
Use when the workload needs 128k context, 4B parameters, and function calling.
Use when the workload needs 128k context, 4B parameters, and function calling.
Use when the workload needs 256k context, 26B parameters, and function calling.
Use when the workload needs 256k context, 26B parameters, and function calling.
Use when the workload needs 256k context, 31B parameters, and function calling.
Use when the workload needs 256k context, 31B parameters, and function calling.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Gemma 4 12B | Use when the workload needs 256k context, 12B parameters, and reasoning. | 2026-06 | 256k context12B parametersreasoning | Current |
| Gemma 4 12B IT | Use when the workload needs 256k context, 12B parameters, and reasoning. | 2026-06 | 256k context12B parametersreasoning | Current |
| Gemma 4 E2B | Use when the workload needs 128k context, 2B parameters, and function calling. | 2026-03 | 128k context2B parametersfunction calling | Current |
| Gemma 4 E2B IT | Use when the workload needs 128k context, 2B parameters, and function calling. | 2026-03 | 128k context2B parametersfunction calling | Current |
| Gemma 4 E4B | Use when the workload needs 128k context, 4B parameters, and function calling. | 2026-03 | 128k context4B parametersfunction calling | Current |
| Gemma 4 E4B IT | Use when the workload needs 128k context, 4B parameters, and function calling. | 2026-03 | 128k context4B parametersfunction calling | Current |
| Gemma 4 26B A4B | Use when the workload needs 256k context, 26B parameters, and function calling. | 2026-03 | 256k context26B parametersfunction calling | Current |
| Gemma 4 26B A4B IT | Use when the workload needs 256k context, 26B parameters, and function calling. | 2026-03 | 256k context26B parametersfunction calling | Current |
| Gemma 4 31B | Use when the workload needs 256k context, 31B parameters, and function calling. | 2026-03 | 256k context31B parametersfunction calling | Current |
| Gemma 4 31B IT | Use when the workload needs 256k context, 31B parameters, and function calling. | 2026-03 | 256k context31B parametersfunction calling | Current |
Release Timeline
2 release groupsSpecifications(10 models)
| Model | Released | Context | Parameters | Vision | Multimodal | Reasoning | Fn Calling | Tool Use | Structured Outputs |
|---|---|---|---|---|---|---|---|---|---|
| Gemma 4 12B | 2026-06 | 256k | 12B | Yes | Yes | Yes | Yes | Yes | Yes |
| Gemma 4 12B IT | 2026-06 | 256k | 12B | Yes | Yes | Yes | Yes | Yes | Yes |
| Gemma 4 E2B | 2026-03 | 128k | 2B | No | Yes | No | Yes | No | No |
| Gemma 4 E2B IT | 2026-03 | 128k | 2B | No | Yes | No | Yes | No | Yes |
| Gemma 4 E4B | 2026-03 | 128k | 4B | No | Yes | No | Yes | No | No |
| Gemma 4 E4B IT | 2026-03 | 128k | 4B | No | Yes | No | Yes | No | Yes |
| Gemma 4 26B A4B | 2026-03 | 256k | 26B | Yes | Yes | No | Yes | No | No |
| Gemma 4 26B A4B IT | 2026-03 | 256k | 26B | Yes | Yes | No | Yes | No | Yes |
| Gemma 4 31B | 2026-03 | 256k | 31B | Yes | Yes | No | Yes | No | No |
| Gemma 4 31B IT | 2026-03 | 256k | 31B | Yes | Yes | No | Yes | No | Yes |
Available From(12 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| Gemma 4 26B A4B IT | OpenRouter | $0.06 | $0.33 | Serverless |
| Gemma 4 26B A4B IT | Cloudflare Workers AI | $0.1 | $0.3 | Serverless |
| Gemma 4 31B IT | OpenRouter | $0.13 | $0.38 | Serverless |
| Gemma 4 26B A4B IT | Vercel AI Gateway | $0.13 | $0.4 | Serverless |
| Gemma 4 26B A4B IT | Novita AI | $0.13 | $0.4 | Serverless |
| Gemma 4 26B A4B IT | NextBit | $0.13 | $0.4 | Serverless |
| Gemma 4 31B | Vercel AI Gateway | $0.14 | $0.4 | Serverless |
| Gemma 4 31B IT | Novita AI | $0.14 | $0.4 | Serverless |
| Gemma 4 26B A4B IT | GCP Vertex AI | $0.15 | $0.6 | Serverless |
| Gemma 4 31B IT | GCP Vertex AI | $0.15 | $0.6 | Serverless |
| Gemma 4 31B IT | Together AI | $0.39 | $0.97 | Serverless |
Popular comparisons in this family
- Claude Haiku 4.5 vs Gemma 4 E2B285
- Claude Sonnet 4.5 vs Gemma 4 26B A4B IT166
- Claude Haiku 4.5 vs Gemma 4 31B IT130
- Gemma 4 12B vs Gemma 4 E4B129
- Claude Sonnet 4.5 vs Gemma 4 E2B118
- Claude Haiku 4.5 vs Gemma 4 26B A4B IT86
- DeepSeek V3 vs Gemma 4 31B IT85
- DeepSeek V3 vs Gemma 4 26B A4B IT58
- Claude Opus 4.5 vs Gemma 4 E2B41
- Claude Opus 4.5 vs Gemma 4 26B A4B IT31
Comparisons
- Gemma 4 12B IT vs Gemma 4 E4B IT
- Gemma 4 12B IT vs Gemma 4 26B A4B IT
- Gemma 4 12B IT vs Gemma 3 12B
- Gemma 4 12B IT vs Phi-4 14B
- Gemma 4 12B IT vs Llama 4 Scout 17B-16E Instruct
Frequently Asked Questions
- What is Gemma 4 used for?
- Gemma 4 is used for multimodal, vision and multimodal work, and reasoning. The family description and listed model capabilities point to those workloads as the best fit.
- How does Gemma 4 compare to T5Gemma?
- Gemma 4 by Google DeepMind is strongest where you need multimodal, while T5Gemma by Google DeepMind is the closest related family to check for agent workflows and tool use. Gemma 4 has 10 listed variants and reaches up to 256k context, so compare the specs and pricing tables before choosing a production model.
- Which Gemma 4 model should I use?
- For the lowest listed input price, start with Gemma 4 26B A4B IT through OpenRouter at $0.06/1M input tokens. For the most capable/latest local choice, evaluate Gemma 4 12B with 256k context and reasoning, tool use, function calling, structured outputs, and multimodal inputs.
Models(10)
Gemma 4 12B
Gemma 4 12B IT
Gemma 4 E2B
Gemma 4 E2B IT
Gemma 4 E4B
Gemma 4 E4B IT
Gemma 4 26B A4B
Gemma 4 26B A4B IT
Gemma 4 31B
Gemma 4 31B IT



