Gemini 4 Models by Google DeepMind
Google DeepMindProprietary
Google DeepMind releases · 51 in the last 12 months · this family litChangelog →
1 model2026
Last refreshed 2026-09-30. Next refresh: weekly.
Details
ResearcherGoogle DeepMind
LicenseProprietary
Commercial useCommercial use: conditional
Models1
Released2026
Capabilities
VisionAll models
MultimodalAll models
ReasoningAll models
Links
WebsiteAbout
Google DeepMind's Gemini 4 frontier model family opened September 30, 2026 with Gemini 4 Argon — a new major generation after the Gemini 3.x line, with a chemical-element code name for the first SKU. First public model gemini-4-argon targets complex long-horizon workflows across coding, enterprise knowledge work, and cybersecurity defense. Own major-version family; do not park under gemini-3.8 or gemini-3. Distinct from Gemma 4 open weights.
Decision facts
- Best fit
- codingvision and multimodal workreasoning
- Capability starting point
- Gemini 4 Argon with reasoning and multimodal inputs
- Lowest tracked input
- Not tracked
- Closest related family
- DiffusionGemma
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
1 in view
Gemini 4 ArgonCurrent
Use when the workload needs reasoning, multimodal inputs, and coding.
2026-09reasoningmultimodal inputscoding
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Gemini 4 Argon | Use when the workload needs reasoning, multimodal inputs, and coding. | 2026-09 | reasoningmultimodal inputscoding | Current |
Release Timeline
1 release group2026-09
1 current
Gemini 4 Argon
Currentreasoningmultimodal inputscoding
Specifications(1 models)
| Model | Released | Vision | Multimodal | Reasoning |
|---|---|---|---|---|
| Gemini 4 Argon | 2026-09 | Yes | Yes | Yes |
