--- title: "Google's Gemini 3.5 Flash Beats Pro While Antigravity 2.0 Kills the CLI" subtitle: "Google I/O 2026 drops Gemini 3.5 Flash, Antigravity 2.0, and a full agentic dev stack — all generally available today" url: "https://alphasignal.ai/news/google-s-gemini-3-5-flash-beats-pro-while-antigravity-2-0-kills-the-cli" author: "AlphaSignal Newsroom" author_url: "https://alphasignal.ai/editorial-team" company: "Google for Developers" company_url: "https://alphasignal.ai/company/googledevs" published: "2026-05-26T19:00:00.000Z" updated: "2026-06-08T19:40:17.523Z" type: "News" topics: - "Agents" - "Development" subtopics: - "Agent Frameworks" - "Code Agents" - "Code Agents" upvotes: 252 read_time: "8 min" source_url: "https://www.linkedin.com/pulse/google-io-2026-developer-recap-game-changing-announcements-nezfc/?trackingId=MhIwN6664cZP2n506CAipg%3D%3D" comments: 0 --- # Google's Gemini 3.5 Flash Beats Pro While Antigravity 2.0 Kills the CLI > Google I/O 2026 drops Gemini 3.5 Flash, Antigravity 2.0, and a full agentic dev stack — all generally available today Google I/O 2026 was not a product showcase. It was a platform declaration. Every major announcement pointed at the same target: turning AI from a tool you query into infrastructure that acts on your behalf. Google is no longer building a chatbot. It's building an agent platform. And at the center of that shift are three developer-facing announcements that landed generally available the same day they were announced. ## A Flash model that beats the old Pro Gemini 3.5 Flash is Google's biggest Flash-tier model launch ever, and it shipped generally available the same day it was announced. As of the keynote, it is the default model in the Gemini app and AI Mode in Google Search worldwide. That alone would be notable. What makes it genuinely surprising is the benchmark story. The headline model announcement at Google I/O 2026 is Gemini 3.5 Flash, and it's a strange one: a Flash-tier model that beats Gemini 3.1 Pro across coding, reasoning, and multimodal benchmarks. That's never happened before in Google's model lineup. Flash models have historically been the cheaper, faster, lower-quality option. That tradeoff is now gone. The numbers back it up. It outperforms Gemini 3.1 Pro on key benchmarks: Terminal-Bench 2.1 at 76.2%, GDPval-AA at 1656 Elo, and MCP Atlas at 83.6%, while leading in multimodal understanding with CharXiv at 84.2%. And it does this while running fast. Google says 3.5 Flash delivers four times the output token generation speed of competing frontier models. There is one area where it still trails. It falls behind Gemini 3.1 Pro on long-context tasks (MRCR v2 at 128k) and Humanity's Last Exam, where raw knowledge depth matters more than agentic capability. For most production workloads, though, that's a narrow tradeoff. 3.5 Flash offers a balance of performance and speed, ideal for tackling long-horizon agentic tasks, often at less than half the cost of comparable models. The model runs at 289 tokens per second with $1.50/$9 pricing (input/output per million tokens). The model ID to use in the API is `gemini-3.5-flash`. The stable API model ID is `gemini-3.5-flash` with no preview suffix, replacing the `gemini-3-flash-preview` identifier used during the preview window. There's also a breaking API change to watch: the `thinking_budget` integer parameter has been replaced with a string enum called `thinking_level`, with values `minimal`, `low`, `medium` (default), and `high`. Critically, the default dropped from high to medium. How was it built? Google DeepMind engineered these models from the ground up using purpose-built AI infrastructure. This unique co-design of the model and hardware allows them to train deeper reasoning capabilities faster and more efficiently with every new generation. Notably, Gemini 3.5 Flash was co-optimized with the Antigravity agent harness, meaning the model and the tooling were developed together , not bolted together after the fact. ## Antigravity 2.0: One platform to replace them all Antigravity 2.0 is now a desktop IDE, a CLI (invoked as `agy`), an SDK, a Managed Agents API tier inside the Gemini API, and an enterprise deployment path , all shipped simultaneously. The company launched its original Antigravity tool last year as a response to agentic coding software such as Cursor. The 2.0 release is a different kind of product: less IDE, more agent runtime. The keynote demo made the ambition clear. On stage, Antigravity 2.0's army of parallel agents built a working operating system core from scratch for under $1,000 in tokens, then ran a live Doom clone coded directly on top of that brand new OS. That's the kind of demo that's hard to forget. Here's what's actually shipping across the five surfaces: - **Desktop app (Antigravity 2.0):** A new standalone desktop application that acts as a central home for agent interaction, allowing you to orchestrate multiple agents to execute tasks in parallel. It features dynamic subagents for parallelized workflows, scheduled tasks for background automation, and ecosystem integrations across Google AI Studio, Android, and Firebase. - **CLI (`agy`):** For those who prefer to stay in the terminal, the Antigravity CLI delivers a lightweight, high-velocity product surface that lets you create new agents instantly without a graphical user interface. - **SDK:** Provides programmatic access to the same agent harness powering Google's products. Optimized for Gemini models, the SDK lets you define custom agent behaviors and host them on your infrastructure of choice. - **Managed Agents API:** With a single API call, the Antigravity agent provisions a remote Linux environment where the agent can reason, plan, and call tools using the harness; execute code and manage files in an isolated sandbox; and browse the web to fetch and process live data. - **Enterprise (Agent Platform):** Google Cloud customers can now access Antigravity through Agent Platform, inheriting Google Cloud's standard data privacy protections and Terms of Service, ensuring customer data stays in your control and agent activity runs within your secure cloud boundary by default. One of the more powerful design choices in Managed Agents is persistent state. Each interaction creates an environment that can be resumed in follow-up calls with all files and state intact, enabling seamless multi-turn sessions without reinitializing context. You can also extend agents without writing orchestration code: define everything in markdown files like `AGENTS.md` and `SKILL.md` and register them as a named agent. The security model inside the CLI is also worth noting. You can spin up specialized subagents to tackle complex workflows, all protected by built-in cross-platform terminal sandboxing, credential masking, and hardened Git policies. ## The Gemini CLI is dead , here's your migration window If you're running production pipelines on Gemini CLI, this is the most urgent item in the entire I/O release. Google Antigravity went from a single IDE to a five-surface agent platform in one release, and simultaneously killed Gemini CLI with a June 18, 2026 shutdown date. Enterprise license and paid API key users can continue to use the Gemini CLI, but consumer and standard developer tiers must migrate. The Antigravity CLI fully replaces the Gemini CLI. For developers migrating, Antigravity CLI preserves the most critical Gemini CLI features: Agent Skills, Hooks, Subagents, and Extensions , the last of which are now rebranded as Antigravity plugins. ## Android in hours, not weeks Google AI Studio now includes native Kotlin support to vibe code Android apps. A new migration agent previewed in Android Studio migrates your app code to a native Kotlin Android app regardless of whether your source is React Native, a web framework, or iOS , turning migrations that would have taken weeks into just hours. The AI Studio mobile app rounds out the workflow. Google is enabling developers to act on their ideas from anywhere that inspiration strikes: explain your vision to AI Studio on your phone and return home to find a full-fledged prototype. From there, you can export the project to Antigravity to launch more agents on your own machine. It's a workflow that makes app development feel less like coding and more like project management, leading a team of Gemini-powered AI agents. Native Android support has been added, allowing developers to build Android apps with just a prompt. Google is also introducing support for the Google Play Console directly in Google AI Studio, enabling developers to publish apps to the test track without leaving the Studio environment. ## The bigger picture: what just shifted The scale context matters here. Google CEO Sundar Pichai revealed the company is now processing more than 3.2 quadrillion tokens per month, a massive increase from 480 trillion tokens per month at I/O 2025. That's a 6x jump in a year, and it's the infrastructure backdrop against which all of these developer tools are being deployed. Google built Antigravity 2.0 using Antigravity 2.0. Every consumer product Google launched at I/O 2026 , Gemini Spark, AI Mode in Search, Daily Brief , runs on the same Antigravity agent harness internally. That's not just a marketing claim. It means the same runtime you're building on is the one Google is betting its own products on. The competitive framing is direct. Antigravity 2.0 is going head to head with Claude Code, Cursor, Windsurf with Devin, OpenAI Codex, and AWS Kiro. The CLI component is a direct competitor to Claude Code and OpenAI's Codex CLI. The difference Google is betting on: deep integration with its own infrastructure stack (Firebase, Android, Cloud Run, Workspace) and a model that was co-optimized with the agent harness from day one. For pricing, Google AI Ultra is a new $100/month tier that includes access to Gemini Spark beta, 20TB of cloud storage, and priority access to new model releases. It offers 5x the usage limits in Antigravity compared to the AI Pro plan. The previous top tier was reduced from $250 to $200/month. Looking ahead: Gemini 3.5 Pro arrives in June 2026. If 3.5 Flash already beats 3.1 Pro, the full 3.5 Pro model should be a significant leap. Build on 3.5 Flash now, and treat 3.5 Pro as your future premium tier when it lands. The [official developer highlights](https://blog.google/innovation-and-ai/technology/developers-tools/google-io-2026-developer-highlights/) from Google, the [Antigravity blog](https://antigravity.google/blog/google-io-2026), and the [Google Cloud I/O recap](https://cloud.google.com/blog/products/ai-machine-learning/innovations-from-google-io-26-on-google-cloud) are the best starting points for getting into the details. There's also a [full developer keynote breakdown](https://developers.googleblog.com/all-the-news-from-the-google-io-2026-developer-keynote/) covering WebMCP, Android CLI, and the rest of the stack. ## Sources - [X post by @googledevs](https://x.com/googledevs/status/2059348847671783625) - [Building the agentic future: Developer highlights from I/O 2026 — Google Blog](https://blog.google/innovation-and-ai/technology/developers-tools/google-io-2026-developer-highlights/) - [Innovations from Google I/O 26 on Google Cloud](https://cloud.google.com/blog/products/ai-machine-learning/innovations-from-google-io-26-on-google-cloud) - [Google Antigravity Blog: Google I/O 2026](https://antigravity.google/blog/google-io-2026) - [All the news from the Google I/O 2026 Developer keynote — Google Developers Blog](https://developers.googleblog.com/all-the-news-from-the-google-io-2026-developer-keynote/) - [Google I/O 2026: Everything Developers Need to Know — AIMadeTools](https://www.aimadetools.com/blog/google-io-2026-developer-announcements/)