# Google DeepMind: what I've logged

> Google DeepMind comes up in 28 posts of my AI log, from September 2025 to October 2026, and each item below links to the primary source. The releases I logged include Gemini 4 Argon for Robotics, Nano Banana 2.1 and EmbeddingGemma 2, filed under frontier, open weights and robotics.

## Items

1. **Gemini 4 Argon: Built to solve your hardest problems** (Google DeepMind, 13 Oct 2026): Google's new frontier model, live in the API, Antigravity and the Gemini app, priced at $2 per million input tokens and $10 per million output tokens at launch. [Primary source](https://deepmind.google/blog/introducing-gemini-4-argon) · [Post](https://heyaliasif.com/ai/2026-10-14-gemini-4-argon-lfm-runai#item-1)
2. **Gemini 4 Argon for Robotics** (Google DeepMind, 13 Oct 2026): Reports 87% on the ERQA embodied reasoning benchmark and shows prompts for robot control, Real2Sim and hindsight labeling. [Primary source](https://deepmind.google/blog/gemini-4-argon-for-robotics) · [Post](https://heyaliasif.com/ai/2026-10-14-gemini-4-argon-lfm-runai#item-2)
3. **Nano Banana 2.1** (Google DeepMind, 5 Oct 2026): Google's image generation and conversational editing model is generally available, with up to 14 reference images and search grounding. [Primary source](https://ai.google.dev/gemini-api/docs/models/gemini-nano-banana-2.1) · [Post](https://heyaliasif.com/ai/2026-10-05-week-in-ai#item-4)
4. **EmbeddingGemma 2** (Google DeepMind, 5 Oct 2026): A 740M open embedding model that maps text, code, images, video and audio into one vector space, across more than 100 languages. [Primary source](https://blog.google/innovation-and-ai/technology/developers-tools/embeddinggemma-2/) · [Post](https://heyaliasif.com/ai/2026-10-05-week-in-ai#item-9)
5. **Gemini 4 Argon** (Google DeepMind, 28 Sept 2026): Google's new frontier model, rolling out first to trusted cyber defenders, with a 1M-token output limit and a stated 77.9% on DeepSWE v1.1. [Primary source](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/) · [Post](https://heyaliasif.com/ai/2026-09-28-week-in-ai#item-1)
6. **SynthID Bio** (Google DeepMind, 28 Sept 2026): Watermarking methods for synthetic biology, so that AI-designed sequences can be traced back to their source. [Primary source](https://deepmind.google/blog/introducing-synthid-bio/) · [Post](https://heyaliasif.com/ai/2026-09-28-week-in-ai#item-8)
7. **Gemini 3.8 Flash TTS and Flash-Lite TTS** (Google DeepMind, 21 Sept 2026): Google's text-to-speech models move to the 3.8 generation, in two sizes. [Primary source](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-text-to-speech/) · [Post](https://heyaliasif.com/ai/2026-09-21-week-in-ai#item-7)
8. **Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking** (Google DeepMind, 14 Sept 2026): Two real-time voice models, one built for scale across 97 languages and one for multi-step work, both reachable through the Gemini API. [Primary source](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-live-gemini-3-8-live-extended-thinking/) · [Post](https://heyaliasif.com/ai/2026-09-14-week-in-ai#item-1)
9. **AlphaGenome Atlas** (Google DeepMind, 7 Sept 2026): A 1-petabyte precomputed map of predicted effects for about 9 billion possible single-letter DNA changes, free for academic use. [Primary source](https://deepmind.google/blog/alphagenome-atlas-a-predictive-map-of-every-possible-dna-letter-change-in-the-human-genome/) · [Post](https://heyaliasif.com/ai/2026-09-07-week-in-ai#item-5)
10. **Gemini 3.8 Flash and Gemini 3.8 Flash Cyber** (Google DeepMind, 31 Aug 2026): Google's most capable Flash model for long-horizon software and agent work, with a separate cyber variant for defenders. [Primary source](https://blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-8-flash-cyber/) · [Post](https://heyaliasif.com/ai/2026-08-31-week-in-ai#item-3)
11. **WeatherNext 3** (Google DeepMind, 31 Aug 2026): DeepMind's most advanced global weather model yet, following the cyclone work from last month. [Primary source](https://blog.google/innovation-and-ai/models-and-research/google-deepmind/introducing-weathernext-3/) · [Post](https://heyaliasif.com/ai/2026-08-31-week-in-ai#item-7)
12. **Gemini Omni 1.1 Flash** (Google DeepMind, 24 Aug 2026): A generally available video generation and editing model with native audio, output up to 4K, scene extension and video references. [Primary source](https://blog.google/innovation-and-ai/technology/developers-tools/build-with-gemini-omni-1-1-flash/) · [Post](https://heyaliasif.com/ai/2026-08-24-week-in-ai#item-1)
13. **Piloting the world's first double-blind AI evaluations** (Google DeepMind, 24 Aug 2026): Outside evaluators tested a Gemini model without seeing its weights, and Google never saw the test prompts, which limits benchmark contamination. [Primary source](https://deepmind.google/blog/piloting-the-worlds-first-double-blind-ai-evaluations/) · [Post](https://heyaliasif.com/ai/2026-08-24-week-in-ai#item-7)
14. **From Atari to EVE Online: 15 years of AI research in games** (Google DeepMind, 17 Aug 2026): Traces DQN through AlphaGo and SIMA 2, and announces a partnership to study long-term memory and multi-agent behavior inside EVE Online. [Primary source](https://deepmind.google/blog/from-atari-to-eve-online-building-on-15-years-of-ai-research-in-games/) · [Post](https://heyaliasif.com/ai/2026-08-17-week-in-ai#item-3)
15. **Gemini 3.7 Flash** (Google DeepMind, 10 Aug 2026): Google's generally available workhorse for coding and agents, with a 1M-token window, 64K output and thinking levels from low to high. [Primary source](https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash/) · [Post](https://heyaliasif.com/ai/2026-08-10-week-in-ai#item-2)
16. **SL2T sign-language-to-text** (Google DeepMind, 10 Aug 2026): A sign-to-text model that runs on pose landmarks on the device, so no raw video leaves the phone, starting with American Sign Language. [Primary source](https://deepmind.google/blog/putting-sign-language-ai-into-users-hands/) · [Post](https://heyaliasif.com/ai/2026-08-10-week-in-ai#item-7)
17. **WeatherNext cyclone forecasting** (Google DeepMind, 3 Aug 2026): A Nature paper reports three-day cyclone forecasts as accurate as older models' two-day ones, and DeepMind is open-sourcing the models. [Primary source](https://deepmind.google/blog/weathernext-ai-model-achieves-breakthrough-in-forecasting-cyclones/) · [Post](https://heyaliasif.com/ai/2026-08-03-week-in-ai#item-4)
18. **Gemini Robotics 2 and Gemini Robotics ER 2** (Google DeepMind, 27 Jul 2026): Three robot models covering whole-body control, embodied reasoning and on-device use, with ER 2 already open to developers in Google AI Studio. [Primary source](https://deepmind.google/blog/gemini-robotics-2-brings-whole-body-intelligence-to-robots/) · [Post](https://heyaliasif.com/ai/2026-07-27-week-in-ai#item-1)
19. **Gemini 3.6 Flash, Gemini 3.5 Flash-Lite and Gemini 3.5 Flash Cyber** (Google DeepMind, 20 Jul 2026): A generally available multimodal Flash model with a 1M-token window at $0.75 in and $3.75 out per million tokens, plus a cheaper Lite tier. [Primary source](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-6-flash-3-5-flash-lite-3-5-flash-cyber/) · [Post](https://heyaliasif.com/ai/2026-07-20-week-in-ai#item-2)
20. **Gemini Omni Flash and Nano Banana 2 Lite** (Google DeepMind, 29 Jun 2026): June 30 brought a video generation and editing model at $0.10 per second and a fast image model at $0.034 per 1K image, both with SynthID. [Primary source](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-omni-flash-nano-banana-2-lite/) · [Post](https://heyaliasif.com/ai/2026-06-29-week-in-ai#item-3)
21. **Computer use in Gemini 3.5 Flash** (Google DeepMind, 22 Jun 2026): June 24 makes computer use a built-in tool in Gemini 3.5 Flash, with optional confirmation steps and prompt-injection stops for enterprises. [Primary source](https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-computer-use-gemini-3-5-flash/) · [Post](https://heyaliasif.com/ai/2026-06-22-week-in-ai#item-4)
22. **Gemini 3.5 Flash** (Google DeepMind, 18 May 2026): First model of the 3.5 family, reported at 76.2% on Terminal-Bench 2.1 and 83.6% on MCP Atlas, and the new default in the Gemini app and AI Mode. [Primary source](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/) · [Post](https://heyaliasif.com/ai/2026-05-18-week-in-ai#item-1)
23. **Gemini Omni** (Google DeepMind, 18 May 2026): A model that creates content from mixed inputs, starting with video. [Primary source](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-omni/) · [Post](https://heyaliasif.com/ai/2026-05-18-week-in-ai#item-2)
24. **Project Genie with Street View** (Google DeepMind, 18 May 2026): Experimental interactive 3D simulations of real places, running in the browser. [Primary source](https://blog.google/innovation-and-ai/models-and-research/google-deepmind/project-genie-expands/) · [Post](https://heyaliasif.com/ai/2026-05-18-week-in-ai#item-5)
25. **Gemini Robotics-ER 1.6** (Google DeepMind, 13 Apr 2026): Better spatial reasoning and multi-view understanding, plus gauge reading developed with Boston Dynamics, available in the Gemini API and AI Studio. [Primary source](https://deepmind.google/blog/gemini-robotics-er-1-6/) · [Post](https://heyaliasif.com/ai/2026-04-13-week-in-ai#item-4)
26. **Gemma 4** (Google DeepMind, 30 Mar 2026): Four sizes (E2B, E4B, a 26B mixture-of-experts and a 31B dense) under Apache 2.0, with the 31B ranked third among open models on the Arena text leaderboard, per Google. [Primary source](https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/) · [Post](https://heyaliasif.com/ai/2026-03-30-week-in-ai#item-1)
27. **Gemini 3.1 Flash Live** (Google DeepMind, 23 Mar 2026): A real-time audio model for voice agents, with lower latency and better function calling. [Primary source](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-flash-live/) · [Post](https://heyaliasif.com/ai/2026-03-23-week-in-ai#item-1)
28. **Gemini Embedding 2** (Google DeepMind, 9 Mar 2026): One embedding model maps text, images, video and audio into the same space, so a single search can span all of them. [Primary source](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-embedding-2/) · [Post](https://heyaliasif.com/ai/2026-03-09-week-in-ai#item-2)
29. **Gemini 3.1 Flash-Lite** (Google DeepMind, 2 Mar 2026): Google's cheapest Gemini 3 tier, built for high-volume work where cost per call is the constraint. [Primary source](https://deepmind.google/models/model-cards/gemini-3-1-flash-lite/) · [Post](https://heyaliasif.com/ai/2026-03-02-week-in-ai#item-2)
30. **Gemini 3.1 Pro** (Google DeepMind, 16 Feb 2026): Google's updated top model, with a model card and a stronger showing on reasoning benchmarks. [Primary source](https://deepmind.google/models/model-cards/gemini-3-1-pro/) · [Post](https://heyaliasif.com/ai/2026-02-16-week-in-ai#item-2)
31. **Gemini 3 Deep Think update** (Google DeepMind, 9 Feb 2026): Google upgraded its extended-reasoning mode, pitched at hard science and engineering problems. [Primary source](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-deep-think/) · [Post](https://heyaliasif.com/ai/2026-02-09-week-in-ai#item-1)
32. **Project Genie** (Google DeepMind, 26 Jan 2026): Genie 3 became something users can try: you describe a world and walk around in it. [Primary source](https://deepmind.google/blog/project-genie-experimenting-with-infinite-interactive-worlds/) · [Post](https://heyaliasif.com/ai/2026-01-26-week-in-ai#item-2)
33. **Nano Banana Pro** (Google DeepMind, 17 Nov 2025): Image generation and editing on Gemini 3, with sharp text rendering and more control over composition. [Primary source](https://blog.google/technology/ai/nano-banana-pro/) · [Post](https://heyaliasif.com/ai/2025-11-17-week-in-ai#item-5)
34. **SIMA 2** (Google DeepMind, 10 Nov 2025): A Gemini-based agent that plays in 3D game worlds, reasons about goals and improves itself. [Primary source](https://deepmind.google/blog/sima-2-an-agent-that-plays-reasons-and-learns-with-you-in-virtual-3d-worlds/) · [Post](https://heyaliasif.com/ai/2025-11-10-week-in-ai#item-3)
35. **Gemini 2.5 Computer Use model** (Google DeepMind, 6 Oct 2025): A Gemini model that clicks, types and scrolls through a real browser through the API. [Primary source](https://blog.google/technology/google-deepmind/gemini-computer-use-model/) · [Post](https://heyaliasif.com/ai/2025-10-06-week-in-ai#item-1)
36. **CodeMender** (Google DeepMind, 6 Oct 2025): An agent that finds and patches security bugs, with human review before anything is merged. [Primary source](https://deepmind.google/discover/blog/introducing-codemender-an-ai-agent-for-code-security/) · [Post](https://heyaliasif.com/ai/2025-10-06-week-in-ai#item-3)
37. **Gemini Robotics 1.5** (Google DeepMind, 22 Sept 2025): A robot model that plans in natural language before acting, paired with an embodied reasoning model that can call tools. [Primary source](https://deepmind.google/discover/blog/gemini-robotics-15-brings-ai-agents-into-the-physical-world/) · [Post](https://heyaliasif.com/ai/2025-09-22-week-in-ai#item-6)

## Posts

- 2026-10-14 [Gemini 4 Argon lands, Run:ai goes open source](https://heyaliasif.com/ai/2026-10-14-gemini-4-argon-lfm-runai)
- 2026-10-05 [Mistral Large 4, Claude Haiku 5.5 and a wave of decision models](https://heyaliasif.com/ai/2026-10-05-week-in-ai)
- 2026-09-28 [Gemini 4 Argon debuts to cyber defenders, Sonnet 5.5, GPT-6.1 Sol](https://heyaliasif.com/ai/2026-09-28-week-in-ai)
- 2026-09-21 [Claude Opus 5.5 and GPT-6 Sol land, Claude finds a new enzyme](https://heyaliasif.com/ai/2026-09-21-week-in-ai)
- 2026-09-14 [Gemini 3.8 Live, open Qwen-Image-2.1 and Anthropic's Accenture pact](https://heyaliasif.com/ai/2026-09-14-week-in-ai)
- 2026-09-07 [DeepSeek V4.1 Flash, Devin's SWE-2 and a petabyte of DNA predictions](https://heyaliasif.com/ai/2026-09-07-week-in-ai)
- 2026-08-31 [Claude Fable 5.1 and GPT-6 Astra both land, plus Gemini 3.8 Flash](https://heyaliasif.com/ai/2026-08-31-week-in-ai)
- 2026-08-24 [Gemini Omni 1.1 Flash video, Tencent's Hy4 and GLM-5.3-Flash open up](https://heyaliasif.com/ai/2026-08-24-week-in-ai)
- 2026-08-17 [A quiet week: Qwen3.8-27B, DeepSeek adds vision, DeepMind on games](https://heyaliasif.com/ai/2026-08-17-week-in-ai)
- 2026-08-10 [Grok 4.6, Gemini 3.7 Flash and GLM-5.3 arrive in four days](https://heyaliasif.com/ai/2026-08-10-week-in-ai)
- 2026-08-03 [Qwen3.8-Max at 2.4T, Motif 3, and WeatherNext goes open](https://heyaliasif.com/ai/2026-08-03-week-in-ai)
- 2026-07-27 [Gemini Robotics 2, plus Korean labs open 700B-class models](https://heyaliasif.com/ai/2026-07-27-week-in-ai)
- 2026-07-20 [Claude Opus 5 arrives, Gemini 3.6 Flash and Solar Open 2 follow](https://heyaliasif.com/ai/2026-07-20-week-in-ai)
- 2026-06-29 [Claude Sonnet 5 at $2 and $10, Fable 5 returns, Nano Banana 2 Lite](https://heyaliasif.com/ai/2026-06-29-week-in-ai)
- 2026-06-22 [Claude Tag in Slack, Sakana Fugu, Mistral OCR 4, Codex Remote](https://heyaliasif.com/ai/2026-06-22-week-in-ai)
- 2026-05-18 [Gemini 3.5 Flash at I/O, Qwen3.7-Max, Cursor Composer 2.5](https://heyaliasif.com/ai/2026-05-18-week-in-ai)
- 2026-04-13 [Claude Opus 4.7, Seedance 2.0 report, Gemini Robotics-ER 1.6](https://heyaliasif.com/ai/2026-04-13-week-in-ai)
- 2026-03-30 [Gemma 4 goes Apache 2.0, Qwen3.6-Plus, GLM-5V-Turbo](https://heyaliasif.com/ai/2026-03-30-week-in-ai)
- 2026-03-23 [Google's TurboQuant, open voice from Mistral and Cohere, Flash Live](https://heyaliasif.com/ai/2026-03-23-week-in-ai)
- 2026-03-09 [Nemotron 3 Super opens up and Gemini Embedding 2 goes multimodal](https://heyaliasif.com/ai/2026-03-09-week-in-ai)
- 2026-03-02 [GPT-5.4 ships, Karpathy's autoresearch, and Qwen3.5 goes small](https://heyaliasif.com/ai/2026-03-02-week-in-ai)
- 2026-02-16 [Qwen3.5, Claude Sonnet 4.6 and Gemini 3.1 Pro in one week](https://heyaliasif.com/ai/2026-02-16-week-in-ai)
- 2026-02-09 [GLM-5, MiniMax M2.5 and Seedance 2.0 arrive before Lunar New Year](https://heyaliasif.com/ai/2026-02-09-week-in-ai)
- 2026-01-26 [Kimi K2.5, DeepMind's Project Genie, and the OpenClaw frenzy](https://heyaliasif.com/ai/2026-01-26-week-in-ai)
- 2025-11-17 [Gemini 3 Pro, Antigravity, Grok 4.1, Nano Banana Pro, SAM 3](https://heyaliasif.com/ai/2025-11-17-week-in-ai)
- 2025-11-10 [GPT-5.1, World Labs' Marble, and Anthropic's espionage report](https://heyaliasif.com/ai/2025-11-10-week-in-ai)
- 2025-10-06 [DevDay apps, Gemini drives a browser, 7M-parameter reasoning](https://heyaliasif.com/ai/2025-10-06-week-in-ai)
- 2025-09-22 [Qwen3-Omni and VL arrive, Gemini Robotics 1.5 learns to think first](https://heyaliasif.com/ai/2025-09-22-week-in-ai)

Canonical: https://heyaliasif.com/ai/topics/google-deepmind