Vendor · Language models · Image models · Video models · Audio models · Specialist models
Google has 66 models in the directory. The strongest is Gemini 4 Carbon: Rank 1 to 5 of 95 in the overall index, with 2 models among the best ten.
Google DeepMind is Google’s AI research unit. It was formed in 2023 by bringing Google Brain and DeepMind together and works on AI systems such as Gemini.
In the directory: 43 language models, 4 image models, 6 video models, 11 audio models and 2 specialist models. 19 of them with open weights that can run locally.
Every model there ever was. Models with a profile lead to every benchmark, price and source.
| Gemini 4 Argon | Oct 13, 2026 | Active | 93Rank 2 to 6 of 95 | $2 / $10 | 82 | 1.05M | 97 |
| Gemini 3.8 Flash | Sep 2, 2026 | Active | 67Rank 14 to 19 of 95 | $0.75 / $3.75 | 133 | 1.05M | 57 |
| Gemini 3.7 Flash | Aug 13, 2026 | Deprecated | 70Rank 14 to 20 of 95 | n/a | 306 | 1.05M | 57 |
| Gemini 3.5 Flash-Lite | Jul 21, 2026 | Active | 3Rank 55 to 68 of 95 | $0.30 / $2.50 | 454 | 1.05M | 41 |
| Gemini 3.6 Flash | Jul 21, 2026 | Active | 42Rank 24 to 38 of 95 | $0.75 / $3.75 | 202 | 1.05M | 39 |
| DiffusionGemma 26B A4B IT | Jun 10, 2026 | Active | n/a | n/a | n/a | 262K | 63 |
| Gemma 4 12B IT | Jun 3, 2026 | Active | n/a | n/a | 118 | 262K | 43 |
| Gemini 3.5 Flash | May 19, 2026 | Deprecated | 43Rank 26 to 38 of 95 | n/a | 195 | 1.05M | 38 |
| Gemini 3.1 Flash-Lite | May 7, 2026 | Deprecated | n/a | $0.25 / $1.50 | 380 | 1.05M | 44 |
| Gemma 4 26B A4B IT | Apr 2, 2026 | Active | n/a | n/a | n/a | 262K | 21 |
| Gemma 4 31B IT | Apr 2, 2026 | Active | n/a | n/a | n/a | 262K | 23 |
| Gemma 4 E2B IT | Apr 2, 2026 | Active | n/a | n/a | n/a | 131K | 46 |
| Gemma 4 E4B IT | Apr 2, 2026 | Active | n/a | n/a | 50 | 131K | 46 |
| Gemma 4 26B A4B | Mar 31, 2026 | Active | n/a | $0.09 / $0.31 | 96 | 262K | 50 |
| Gemma 4 31B | Mar 31, 2026 | Active | n/a | $0.14 / $0.40 | 35 | 262K | 53 |
| Gemini 3.1 Pro Preview | Feb 19, 2026 | Preview | 26Rank 36 to 45 of 95 | $2 / $12 | 145 | 1.05M | 50 |
| Gemini 3 Flash Preview | December 2025 | Preview | 9Rank 45 to 58 of 95 | $0.50 / $3 | 206 | 1.05M | 48 |
| Gemini 3 Prono profile | November 2025 | Retired | n/a | n/a | n/a | n/a | n/a |
| Gemini 2.5 Flash-Lite | Jul 22, 2025 | Active | n/a | $0.10 / $0.40 | 298 | 1.05M | 60 |
| Gemma 3n E2B IT | Jun 26, 2025 | Active | n/a | n/a | n/a | 33K | 51 |
| Gemma 3n E4B IT | Jun 26, 2025 | Active | n/a | n/a | n/a | 33K | 45 |
| Gemini 2.5 Flash | Jun 17, 2025 | Active | n/a | $0.30 / $2.50 | 206 | 1.05M | 56 |
| Gemini 2.5 Pro | Jun 17, 2025 | Active | below 1Rank 64 to 78 of 95 | $1.25 / $10 | 116 | 1.05M | 40 |
| Gemma 3 12B IT | Mar 12, 2025 | Active | n/a | n/a | n/a | 131K | 31 |
| Gemma 3 4B IT | Mar 12, 2025 | Active | n/a | n/a | n/a | 131K | 31 |
| Gemma 3 12B | Mar 10, 2025 | Active | n/a | $0.05 / $0.15 | n/a | 128K | 33 |
| Gemma 3 27B | Mar 10, 2025 | Active | n/a | $0.08 / $0.35 | n/a | 128K | 37 |
| Gemma 3 4B | Mar 10, 2025 | Active | n/a | $0.04 / $0.05 | n/a | 128K | 23 |
| Gemini 2.0 Flash-Liteno profile | Feb 5, 2025 | Retired | n/a | n/a | n/a | n/a | n/a |
| Gemini 2.0 Flashno profile | December 2024 | Retired | n/a | n/a | n/a | n/a | n/a |
| Gemma 2 2B IT | Jul 31, 2024 | Active | n/a | n/a | n/a | 8K | 56 |
| Gemma 2 27B | Jun 27, 2024 | Active | n/a | $0.65 / $0.65 | n/a | 8K | 31 |
| Gemma 2 9Bno profile | June 2024 | No status data | n/a | n/a | n/a | n/a | n/a |
| Gemini 1.5 Flashno profile | May 14, 2024 | Retired | n/a | n/a | n/a | n/a | n/a |
| Gemini 1.0 Ultrano profile | Feb 8, 2024 | Retired | n/a | n/a | n/a | n/a | n/a |
| Gemini 1.5 Prono profile | February 2024 | Retired | n/a | n/a | n/a | n/a | n/a |
| Gemini 1.0 Prono profile | Dec 6, 2023 | Retired | n/a | n/a | n/a | n/a | n/a |
| Gemini Nano 1no profile | December 2023 | Active | n/a | n/a | n/a | n/a | n/a |
| Gemini Nano 2no profile | December 2023 | Active | n/a | n/a | n/a | n/a | n/a |
| PaLM 2no profile | May 2023 | No status data | n/a | n/a | n/a | n/a | n/a |
| PaLMno profile | April 2022 | No status data | n/a | n/a | n/a | n/a | n/a |
| LaMDAno profile | January 2022 | No status data | n/a | n/a | n/a | n/a | n/a |
| Gemini 4 Carbon | n/a | Restricted early access | 96Rank 1 to 5 of 95 | n/a | 100 | 1M | below 1 |
| Nano Banana 2.1 | Oct 6, 2026 | Active | n/a | 8 |
| Nano Banana 2 Lite | June 2026 | Active | $0.034 per image | 28 |
| Nano Banana 2 | February 2026 | Active | $0.067 per image | 64 |
| Nano Banana Pro | November 2025 | Active | $0.13 per image | 53 |
| Gemini Omni 1.1 Flash | Aug 27, 2026 | Active | $6.08 per minute | 33 |
| Gemini Omni Flash | June 2026 | Discontinued | $6 per minute | 43 |
| Veo 3.1 Lite Preview | Mar 31, 2026 | Preview | n/a | below 1 |
| Veo 3.1 | Oct 15, 2025 | Active | $24 per minute | 50 |
| Veo 3.1 Fast | Oct 15, 2025 | Active | $6 per minute | 24 |
| Genie 4 | n/a | Research preview for trusted testers | n/a | below 1 |
| Gemini 3.8 Flash TTS | Sep 22, 2026 | Active | n/a | below 1 |
| Gemini 3.8 Flash-Lite TTS | Sep 22, 2026 | Active | n/a | below 1 |
| Gemini 3.1 Flash TTS | n/a | Preview | n/a | 44 |
| Gemini 3.5 Live Translate | n/a | Preview | n/a | 46 |
| Gemini 3.5 Transcribe | n/a | Active | n/a | 52 |
| Gemini 3.5 Transcribe Live | n/a | Active | n/a | 35 |
| Gemini 3.8 Live | n/a | Active | n/a | below 1 |
| Gemini 3.8 Live Extended Thinking | n/a | Active | n/a | below 1 |
| Lyria 3 Clip | n/a | Preview | n/a | 26 |
| Lyria 3.5 | n/a | Active | n/a | 47 |
| Lyria Realtime | n/a | Preview | n/a | 19 |
| EmbeddingGemma 2 | Oct 6, 2026 | Open weights | n/a | 63 |
| Gemini 3.8 Flash Cyber | Sep 2, 2026 | Limited access | n/a | below 1 |
Google names its models by task. Gemini is the family of language models behind the Gemini app and the Gemini API. Within a generation, Pro covers the most demanding tasks, Flash the fast middle tier, and Flash-Lite the cheapest tier, for example for document parsing. Gemini 4 Argon is the first generally available model of the fourth generation, while Gemini 4 Carbon is open to selected partners first.
Gemma are Google's open models whose weights you can download and run locally, Gemma 4 for example under Apache 2.0. Nano Banana creates and edits images, Veo creates videos, and Lyria makes music. On top come models for speech output (TTS), transcription, and live conversations.
Language model · Oct 13, 2026
Google announced Gemini 4 Argon on September 30, 2026 and made it generally available through the Gemini API and Google AI Studio on October 13. Its context window holds 1,048,576 tokens shared across input, thinking, and output.
Specialist model · Oct 6, 2026
EmbeddingGemma 2 extends Google's EmbeddingGemma with shared representations for text, code, images, video, and audio. Its predecessor handled text; the new version maps several modalities into one 768-dimensional vector space.
Image model · Oct 6, 2026
Nano Banana 2.1 improves prompt adherence, character consistency across editing turns, and text rendering over Nano Banana 2, according to Google. A workflow can use up to 14 reference images.
Audio model · Sep 22, 2026
Gemini 3.8 Flash TTS is Google's speech model for nuanced delivery, regional dialects, and long-form multi-turn output.
Audio model · Sep 22, 2026
Gemini 3.8 Flash-Lite TTS is Google's speech model tuned for throughput and cost efficiency in production workloads and real-time agents.
Language model · Sep 2, 2026
Gemini 3.8 Flash improves over Gemini 3.7 Flash mainly on longer software-engineering tasks. Google’s DeepSWE v1.1 result rises from 65.3% to 73.7%. On Terminal-Bench 2.1, Google reports a gain from 85.8% to 89.4%. These results come from the provider’s evaluations.
Video model · Aug 27, 2026
New in Gemini Omni 1.1 Flash are the stable Omni video API and the documented succession to Gemini Omni Flash. Text, images, and videos up to 10 seconds feed generation, editing, or extension. Output ranges from 3 to 10 seconds, 360p through 4K, at 24 frames per second.
Language model · Jul 21, 2026
Gemini 3.5 Flash-Lite is the stable successor to Gemini 3.1 Flash-Lite. Google names document parsing and subagent tasks as focus areas.
Language model · Jul 21, 2026
Google names code generation and spatial reasoning as focus areas for Gemini 3.6 Flash. Its model card lists Gemini 3.5 Flash as the foundation and Gemini 3.7 Flash as the next direct step.
Language model · Jun 10, 2026
DiffusionGemma 26B A4B IT does not generate text token by token but through discrete diffusion: blocks of tokens are denoised step by step in parallel. Google DeepMind aims to raise decoding speed this way.
Language model · Jun 3, 2026
Google Gemma 4 12B IT is distinguished by the 12B Unified variant’s direct multimodal path without separate audio or vision encoders. Released on June 3, 2026, the checkpoint accepts text, image, audio, and video input, generates text, and uses Apache 2.0 weights.
Image model · June 2026
Nano Banana 2 Lite uses Gemini 3.1 Flash Lite Image, 1K output, and Google’s stated priority on speed and scale.
28 more current models are in the model list above.
Derived from the overall index, the API price and the measured speed of the current models.
The strongest choice
Gemini 4 Carbon
The vendor's highest overall index
The fastest
Gemini 3.5 Flash-Lite
454 tokens/s, a typical model manages 98
The 20 best language models, with the models from Google highlighted. “All” and the filter show every model in the index.
20 of 95 models
Provider
Weights
Price (blended per 1M tokens)
Released
Anthropic
OpenAI
Meta
Moonshot AI
DeepSeek
xAI
Z.ai
Xiaomi
Alibaba
StepFun
Mistral AI
MiniMax
Thinking Machines Lab
Inception
| Vendor | Models | Model types | Best language model | Top 10 | Open weights | New in 2026 |
|---|---|---|---|---|---|---|
| Anthropic | 33 | 33 language models | Claude Opus 5.5 (Rank 1 to 3) | 4 | 0 | 13 |
| OpenAI | 63 | 46 language models, 4 image models, 2 video models, 9 audio models, 2 specialist models | GPT-6 Astra (Rank 4 to 8) | 2 | 4 | 27 |
| 66 | 43 language models, 4 image models, 6 video models, 11 audio models, 2 specialist models | Gemini 4 Carbon (Rank 1 to 5) | 2 | 19 | 26 | |
| Meta | 32 | 29 language models, 1 image model, 1 video model, 1 specialist model | Muse Wonder Full (Rank 6 to 9) | 2 | 26 | 7 |
| DeepSeek | 23 | 23 language models | DeepSeek V4.1 Flash 1017 (Rank 13 to 20) | 0 | 23 | 9 |
| Moonshot AI | 12 | 12 language models | Kimi K3.1 Pro (Rank 10 to 13) | 0 | 10 | 7 |
| xAI | 16 | 12 language models, 2 image models, 2 video models | Grok 4.7 (Rank 18 to 24) | 0 | 1 | 12 |
By release date. Grey: retired models.
26 releases since January 2026, by release date.
Prices per 1M tokens for the current language models, according to Google or the provider linked in each model profile.
| Current model | Input | Output | Context |
|---|---|---|---|
| Gemini 4 Argon | $2 | $10 | 1.05M |
| Gemini 3.8 Flash | $0.75 | $3.75 | 1.05M |
| Gemini 3.6 Flash | $0.75 | $3.75 | 1.05M |
| Gemini 3.1 Pro PreviewPreview | $2 | $12 | 1.05M |
| Gemini 3 Flash PreviewPreview | $0.50 | $3 | 1.05M |
| Gemini 3.5 Flash-Lite | $0.30 | $2.50 | 1.05M |
| Gemini 2.5 Pro | $1.25 | $10 | 1.05M |
| Gemma 3 4B | $0.04 | $0.05 | 128K |
| Gemma 3 12B | $0.05 | $0.15 | 128K |
| Gemma 3 27B | $0.08 | $0.35 | 128K |
| Gemma 4 26B A4B | $0.09 | $0.31 | 262K |
| Gemini 2.5 Flash-Lite | $0.10 | $0.40 | 1.05M |
| Gemma 4 31B | $0.14 | $0.40 | 262K |
| Gemini 3.1 Flash-LiteDeprecated | $0.25 | $1.50 | 1.05M |
| Gemini 2.5 Flash | $0.30 | $2.50 | 1.05M |
| Gemma 2 27B | $0.65 | $0.65 | 8K |
| Current model | Price |
|---|---|
| Gemini Omni 1.1 Flash | $6.08 per minute |
| Nano Banana 2 Lite | $0.034 per image |
| Nano Banana 2 | $0.067 per image |
| Nano Banana Pro | $0.13 per image |
| Veo 3.1 | $24 per minute |
| Veo 3.1 Fast | $6 per minute |
Monthly prices in Germany incl. VAT; “≈” is converted from US dollars. The vendor's models are often also part of other providers' plans, see the AI subscription comparison.
| Plan | Per month | Note |
|---|---|---|
| Google AI Free | free | with credits |
| Jules | free | |
| Google AI Plus | €4.99 | with credits |
| Google AI Pro | €21.99 | with credits |
| Gemini Code Assist Standard | ≈ €24.29 | per seat, Team |
| Gemini Code Assist Enterprise | ≈ €57.52 | per seat, Enterprise |
| Google AI Ultra 5x | €99.99 | with credits |
| Google AI Ultra 20x | €219.99 | with credits |
The directory lists 66 Gemini models from Google: 43 language models, 4 image models, 6 video models, 11 audio models and 2 specialist models. It counts every model Google ever released, including retired ones; 53 of them have their own profile. 42 are currently available, the newest being Gemini 4 Argon, EmbeddingGemma 2 and Nano Banana 2.1.
Gemini 4 Carbon. It is at rank 1 to 5 of 95 in the overall index, which combines independent benchmarks from several sources.
Gemini 4 Argon, released on Oct 13, 2026. The timeline below shows every release by date.
Through the API per 1M tokens (input / output): Gemma 3 4B $0.04 / $0.05. The prices of all 16 current language models are in the price table. As a subscription, Gemini starts at €4.99 per month incl. VAT (Google AI Plus).
Yes, with the Google AI Free plan. The cheapest paid plan is Google AI Plus at €4.99 per month incl. VAT.
Google names its models by task. Gemini is the family of language models behind the Gemini app and the Gemini API. Within a generation, Pro covers the most demanding tasks, Flash the fast middle tier, and Flash-Lite the cheapest tier, for example for document parsing. Gemini 4 Argon is the first generally available model of the fourth generation, while Gemini 4 Carbon is open to selected partners first. Gemma are Google's open models whose weights you can download and run locally, Gemma 4 for example under Apache 2.0. Nano Banana creates and edits images, Veo creates videos, and Lyria makes music. On top come models for speech output (TTS), transcription, and live conversations.
In the overall index, yes. Gemini 4 Carbon (rank 1 to 5 of 95) is ahead of the best OpenAI model, GPT-6 Astra (rank 4 to 8 of 95). Single tasks can differ.
Yes, 19 of the 66 models have open weights. Each model profile lists how much memory it needs.
Gemini 3.5 Flash-Lite with 454 tokens per second according to Artificial Analysis. A typical model manages 98.
Up to 1.05 M tokens (≈ 1,570 A4 pages), which 10 of the current models reach, including Gemini 4 Argon. Every model's value is in the model list.
Data as of Nov 1, 2026