Skip to content

How Google Antigravity’s scores are calculated

The full audit trail, recomputed from the verdict data at build time through the same code that produced the leaderboard: verdict × quality × story weight per cell, cells sum to dimension scores, dimensions blend into the PA Score. Every number on the product page is reproducible from this page alone; for why the formula looks like this, see the methodology.

verdict factors: full ×1.0 · partial ×0.6 · disputed ×0.3 · none ×0.0 · n/a excluded from both sides · cell points = weight × quality × factor · cell max = weight × 10

PA Score37/100

Agent-ready 55.1 × 0.30 = 16.53

API quality 2.4 × 0.20 = 0.48

Openness 4.8 × 0.20 = 0.96

Built-in AI 70.2 × 0.15 = 10.53

Automation 55.3 × 0.15 = 8.29

(16.53 + 0.48 + 0.96 + 10.53 + 8.29) ÷ (0.30 + 0.20 + 0.20 + 0.15 + 0.15) = 36.80 ÷ 1.00 = 36.8

Scores are stored to 1 decimal; the product page’s pills round to whole numbers for display. Each dimension below shows the stories, verdicts, and cited evidence behind its number.

Agent-ready55.1/100×0.30 of the PA blend

Outside-in: can YOUR agent reach and drive this product — API, MCP, CLI, headless runs, agent docs.

Point an agent at llms.txt or agent-oriented docsweight 2

2 (weight) × 9 (quality) × 1.0 (full) = 18.0 of 20 max

  • [probe] https://antigravity.google/llms.txtPROBE llms.txt: HTTP 200 at https://antigravity.google/llms.txt # Google Antigravity > Google Antigravity is an advanced agentic coding platform and CLI designed for orchestrating aut
  • [probe] https://antigravity.google/docs/getting-started.mdPROBE docs-md: HTTP 200 at https://antigravity.google/docs/getting-started.md # Getting Started with Antigravity 2.0 ### Download Visit [antigravity.google/download](https://antigravity.google/dow
  • [claimed-docs] https://antigravity.google/docs/getting-startedVisit antigravity.google/download to download Google Antigravity 2.0. Select your operating system below

Run the product headlessly / in CI for automationweight 2

2 (weight) × 7 (quality) × 1.0 (full) = 14.0 of 20 max

  • [claimed-docs] https://antigravity.google/docs/cli/headlessRun Antigravity CLI non-interactively to script agent tasks, integrate with CI pipelines, and capture machine-readable output.
  • [claimed-docs] https://antigravity.google/product/antigravity-2Automate routine checks with Scheduled Tasks, simply define a cron schedule and the agents start and run autonomously in the background.
  • [claimed-docs] https://antigravity.google/product/antigravity-cliEdit, orchestrate, and build all in natural language. Tell your agents what you need, and they’ll work on getting it done.
  • [claimed-docs] https://antigravity.google/product/antigravity-2Create or download fully customizable skills to further your agent’s autonomy and transform how you get work done.

Plug MCP servers into this product so it can use their toolsweight 3

3 (weight) × 8 (quality) × 1.0 (full) = 24.0 of 30 max

  • [claimed-docs] https://antigravity.google/docs/mcpMCP lets Antigravity fetch structured context directly or execute safe actions on your behalf when needed.
  • [claimed-docs] https://antigravity.google/docs/mcplets AI agents and editors securely connect to local developer tools, databases, file parsers, and external remote APIs
  • [claimed-docs] https://antigravity.google/product/antigravity-cliAccess plugins, MCP, skills, and hooks configurations instantly via slash commands, quickly enhancing your workflow.
  • [claimed-docs] https://antigravity.google/product/antigravity-sdkLayer custom Python callables, Model Context Protocol (MCP) servers, and reusable agent skills over our built-in filesystem and terminal tools under a single pipeline.
  • [claimed-docs] https://antigravity.google/docs/pluginsPlugins are namespaced bundles that allow you to extend Antigravity’s capabilities by grouping skills, rules, MCP servers, and hooks into a single package.

Connect an agent via an official MCP serverweight 3

n/a — not applicable to this product: excluded from numerator and denominator

  • [claimed-docs] https://antigravity.google/product/antigravity-cliAccess plugins, MCP, skills, and hooks configurations instantly via slash commands, quickly enhancing your workflow.
  • [claimed-docs] https://antigravity.google/product/antigravity-sdkLayer custom Python callables, Model Context Protocol (MCP) servers, and reusable agent skills over our built-in filesystem and terminal tools under a single pipeline.
  • [claimed-docs] https://antigravity.google/docs/mcpMCP lets Antigravity fetch structured context directly or execute safe actions on your behalf when needed.
  • [claimed-docs] https://antigravity.google/docs/mcplets AI agents and editors securely connect to local developer tools, databases, file parsers, and external remote APIs

Use an official CLIweight 2

2 (weight) × 8 (quality) × 1.0 (full) = 16.0 of 20 max

  • [claimed-docs] https://antigravity.google/product/antigravity-cliEdit, orchestrate, and build all in natural language. Tell your agents what you need, and they’ll work on getting it done.
  • [claimed-docs] https://antigravity.google/product/antigravity-cliHave multiple agents working in parallel, so larger tasks get tackled faster.
  • [claimed-docs] https://antigravity.google/product/antigravity-cliNavigate your entire workflow via standard terminal shortcuts: adjust permissions, themes, and preferences via /config and type /keybindings to customize every shortcut.
  • [claimed-docs] https://antigravity.google/product/antigravity-cliAccess plugins, MCP, skills, and hooks configurations instantly via slash commands, quickly enhancing your workflow.
  • [claimed-docs] https://antigravity.google/docs/cli/gcli-migrationthe CLI automatically detects your existing profiles. An interactive checklist prompts you to choose which assets to migrate
  • [claimed-docs] https://antigravity.google/docs/cli/sandboxSensitive files like ~/.ssh and .env are blocked, anything not explicitly mounted is invisible inside the sandbox
  • [claimed-docs] https://antigravity.google/docs/cli/vim-editor-modeVim editor mode replaces the editing model in every multi-line input surface of the CLI
  • [claimed-docs] https://antigravity.google/docs/cli/headlessRun Antigravity CLI non-interactively to script agent tasks, integrate with CI pipelines, and capture machine-readable output.
  • [community] https://news.ycombinator.com/item?id=48222529I much prefer using Gemini CLI in combination with vscode. It works like a charm. Now, I'll do the same with Antigravity CLI and vscode. It works fine.
  • [probe] https://antigravity.google/product/antigravity-cliofficial CLI documented at https://antigravity.google/product/antigravity-cli

Drive the product through a documented public APIweight 3

3 (weight) × 6 (quality) × 0.6 (partial) = 10.8 of 30 max

  • [claimed-docs] https://antigravity.google/product/antigravity-sdkThe Agent SDK gives you the same tools, agent loop, and context management that power Google Antigravity, programmable in Python.
  • [claimed-docs] https://antigravity.google/product/antigravity-sdkLayer custom Python callables, Model Context Protocol (MCP) servers, and reusable agent skills over our built-in filesystem and terminal tools under a single pipeline.
  • [claimed-docs] https://antigravity.google/product/antigravity-sdkBuild AI agents that autonomously read files, run commands, edit code, and more.
  • [claimed-docs] https://antigravity.google/docs/cli/headlessRun Antigravity CLI non-interactively to script agent tasks, integrate with CI pipelines, and capture machine-readable output.
  • [probe] https://antigravity.google/openapi.jsonPROBE openapi: all candidate paths 404 (https://antigravity.google/openapi.json, https://antigravity.google/swagger.json, https://antigravity.google/api/openapi.json, https://antigravity.google/.well-known/openapi.json)

Issue scoped/least-privilege API credentials for an agentweight 2

2 (weight) × 4 (quality) × 0.3 (disputed) = 2.4 of 20 max

  • [claimed-docs] https://antigravity.google/docs/permissionsPermissions are evaluated across three distinct access lists: Deny... Ask... Allow
  • [claimed-docs] https://antigravity.google/docs/permissionsPermissions are evaluated across three distinct access lists: Deny...Ask...Allow
  • [claimed-docs] https://antigravity.google/docs/cli/sandboxSensitive files like ~/.ssh and .env are blocked, anything not explicitly mounted is invisible inside the sandbox
  • [community] https://news.ycombinator.com/item?id=46048996Google Antigravity exfiltrates data via indirect prompt injection attack: Gemini is not supposed to have access to .env files with default setting, but bypasses its own setting to get access and exfiltrate that data.
  • [community] https://news.ycombinator.com/item?id=46048996The default Allowlist provided with Antigravity includes 'webhook.site', which was used as an exfiltration vector for secrets.

Build against official SDKsweight 2

2 (weight) × 7 (quality) × 1.0 (full) = 14.0 of 20 max

  • [claimed-docs] https://antigravity.google/product/antigravity-sdkThe Agent SDK gives you the same tools, agent loop, and context management that power Google Antigravity, programmable in Python.
  • [claimed-docs] https://antigravity.google/product/antigravity-sdkLayer custom Python callables, Model Context Protocol (MCP) servers, and reusable agent skills over our built-in filesystem and terminal tools under a single pipeline.
  • [claimed-docs] https://antigravity.google/product/antigravity-sdkPass rich multimedia file attachments (images, videos, audio, and documents) to the agent alongside textual instruction prompt lists.
  • [claimed-docs] https://antigravity.google/product/antigravity-sdkBuild AI agents that autonomously read files, run commands, edit code, and more.
  • [probe] https://antigravity.google/openapi.jsonPROBE openapi: all candidate paths 404 (https://antigravity.google/openapi.json, https://antigravity.google/swagger.json, https://antigravity.google/api/openapi.json, https://antigravity.google/.well-known/openapi.json)

Subscribe to events via webhooksweight 2

2 (weight) × 0 (quality) × 0.0 (none) = 0.0 of 20 max

  • [probe] https://antigravity.google/openapi.jsonPROBE openapi: all candidate paths 404 (https://antigravity.google/openapi.json, https://antigravity.google/swagger.json, https://antigravity.google/api/openapi.json, https://antigravity.google/.well-known/openapi.json)

Agent-ready = 99.2 ÷ 180 × 100 = 55.1

API quality2.4/100×0.20 of the PA blend

The programmable surface once an agent is there — machine-readable spec, interactive docs, sandbox, versioning discipline.

Explore an interactive API reference with runnable examplesweight 2

n/a — not applicable to this product: excluded from numerator and denominator

  • [probe] https://antigravity.google/openapi.jsonPROBE openapi: all candidate paths 404 (https://antigravity.google/openapi.json, https://antigravity.google/swagger.json, https://antigravity.google/api/openapi.json, https://antigravity.google/.well-known/openapi.json)

Download a machine-readable API spec (OpenAPI or equivalent)weight 2

2 (weight) × 0 (quality) × 0.0 (none) = 0.0 of 20 max

  • [probe] https://antigravity.google/openapi.jsonPROBE openapi: all candidate paths 404 (https://antigravity.google/openapi.json, https://antigravity.google/swagger.json, https://antigravity.google/api/openapi.json, https://antigravity.google/.well-known/openapi.json)

Test against a sandbox environment without touching production dataweight 1

1 (weight) × 4 (quality) × 0.3 (disputed) = 1.2 of 10 max

  • [claimed-docs] https://antigravity.google/docs/cli/sandboxSensitive files like ~/.ssh and .env are blocked, anything not explicitly mounted is invisible inside the sandbox
  • [community] https://news.ycombinator.com/item?id=46048996Google Antigravity exfiltrates data via indirect prompt injection attack: Gemini is not supposed to have access to .env files with default setting, but bypasses its own setting to get access and exfiltrate that data.
  • [community] https://news.ycombinator.com/item?id=46048996The default Allowlist provided with Antigravity includes 'webhook.site', which was used as an exfiltration vector for secrets.
  • [community] https://news.ycombinator.com/item?id=46048996Antigravity was also vulnerable to the classic Markdown image exfiltration bug, reported a few days prior and flagged as 'intended behavior'.
  • [community] https://news.ycombinator.com/item?id=46103532Google Antigravity just deleted the contents of whole drive - came down to commanding a deletion of a 'directory with space in the name' without quoting, hunting for the word match ending space which was the D:\ component.

Rely on versioned APIs with a documented deprecation policyweight 2

2 (weight) × 0 (quality) × 0.0 (none) = 0.0 of 20 max

  • [probe] https://antigravity.google/openapi.jsonPROBE openapi: all candidate paths 404 (https://antigravity.google/openapi.json, https://antigravity.google/swagger.json, https://antigravity.google/api/openapi.json, https://antigravity.google/.well-known/openapi.json)

API quality = 1.2 ÷ 50 × 100 = 2.4

Openness4.8/100×0.20 of the PA blend

Can you leave, inspect, or self-host — data export, open source, portability.

Do everything through the API that I can do in the UIweight 2

2 (weight) × 4 (quality) × 0.6 (partial) = 4.8 of 20 max

  • [claimed-docs] https://antigravity.google/product/antigravity-sdkThe Agent SDK gives you the same tools, agent loop, and context management that power Google Antigravity, programmable in Python.
  • [claimed-docs] https://antigravity.google/docs/cli/headlessRun Antigravity CLI non-interactively to script agent tasks, integrate with CI pipelines, and capture machine-readable output.
  • [claimed-docs] https://antigravity.google/product/antigravity-sdkLayer custom Python callables, Model Context Protocol (MCP) servers, and reusable agent skills over our built-in filesystem and terminal tools under a single pipeline.
  • [probe] https://antigravity.google/openapi.jsonPROBE openapi: all candidate paths 404 (https://antigravity.google/openapi.json, https://antigravity.google/swagger.json, https://antigravity.google/api/openapi.json, https://antigravity.google/.well-known/openapi.json)

Export all of my data in open formats and leaveweight 3

3 (weight) × 0 (quality) × 0.0 (none) = 0.0 of 30 max

no evidence cited — the verdict rests on absence of evidence, re-checked on refresh

Read the product's source under an open licenseweight 2

2 (weight) × 0 (quality) × 0.0 (none) = 0.0 of 20 max

no evidence cited — the verdict rests on absence of evidence, re-checked on refresh

Self-host the core productweight 3

3 (weight) × 0 (quality) × 0.0 (none) = 0.0 of 30 max

  • [claimed-docs] https://antigravity.google/docs/getting-startedVisit antigravity.google/download to download Google Antigravity 2.0. Select your operating system below
  • [claimed-docs] https://antigravity.google/docs/overviewAntigravity 2.0 serves as your AI agents’ central command center, providing a unified platform to launch, monitor, and orchestrate their activities.
  • [community] https://news.ycombinator.com/item?id=49548452don't buy Google AI subscription before you confirm you have 'Antigravity entitlement'... You can have verified account, bank card added, active AI Pro subscription, and still you won't be able to use even the free tier of Antigravity.
  • [community] https://news.ycombinator.com/item?id=49548452Google Antigravity TOS: 3rd party usage can get Google account suspended. My friend got a ban by using Pi agent with Antigravity. They un-banned him later, but that was enough to stay away.

Openness = 4.8 ÷ 100 × 100 = 4.8

Built-in AI70.2/100×0.15 of the PA blend

Inside-out: how agentic the product itself is for its users — built-in assistants, autonomous features.

Get AI-generated insights and suggestions from my data inside the productweight 2

2 (weight) × 6 (quality) × 0.6 (partial) = 7.2 of 20 max

  • [claimed-docs] https://antigravity.google/product/antigravity-ideGoogle Antigravity's Editor view offers tab autocompletion, natural language code commands, and a configurable, and context-aware configurable agent.
  • [claimed-docs] https://antigravity.google/docs/artifact-reviewPlanning Mode: The agent plans thoroughly before executing tasks... produces structured implementation plans called Artifacts
  • [claimed-docs] https://antigravity.google/docs/artifactsArtifacts include rich markdown plans (Implementation Plans), code diffs, architecture diagrams, images, and browser recordings.
  • [claimed-docs] https://antigravity.google/product/antigravity-ideoffers tab autocompletion, natural language code commands, and a configurable, and context-aware configurable agent
  • [community] https://news.ycombinator.com/item?id=48222529It's not even good, honestly. I was using it for couple weeks before dropping that 2 months ago. The model was not good and slow, the harness was not good, the IDE was subpar vscode clone.

Set up automations that run autonomously in the backgroundweight 2

2 (weight) × 8 (quality) × 1.0 (full) = 16.0 of 20 max

  • [claimed-docs] https://antigravity.google/product/antigravity-2Automate routine checks with Scheduled Tasks, simply define a cron schedule and the agents start and run autonomously in the background.
  • [claimed-docs] https://antigravity.google/docs/featuresWorktree support: Projects natively support Git worktrees, allowing agents to operate in isolated background folders.
  • [claimed-docs] https://antigravity.google/docs/featuresusers can schedule messages to be sent to their agents while they’re away
  • [claimed-docs] https://antigravity.google/docs/cli/headlessRun Antigravity CLI non-interactively to script agent tasks, integrate with CI pipelines, and capture machine-readable output.
  • [claimed-docs] https://antigravity.google/product/antigravity-ideAble to autonomously operate across your editor, terminal, and browser.

Delegate tasks to a built-in AI assistant inside the productweight 3

3 (weight) × 8 (quality) × 1.0 (full) = 24.0 of 30 max

  • [claimed-docs] https://antigravity.google/product/antigravity-ideAble to autonomously operate across your editor, terminal, and browser.
  • [claimed-docs] https://antigravity.google/product/antigravity-cliEdit, orchestrate, and build all in natural language. Tell your agents what you need, and they’ll work on getting it done.
  • [claimed-docs] https://antigravity.google/docs/subagentsan agent can delegate tasks—such as running tests or performing extensive codebase searches—to dedicated subagents.
  • [claimed-docs] https://antigravity.google/docs/subagentsAntigravity comes pre-packaged with several specialized subagents out of the box
  • [claimed-docs] https://antigravity.google/docs/overviewAntigravity 2.0 serves as your AI agents’ central command center, providing a unified platform to launch, monitor, and orchestrate their activities.
  • [community] https://news.ycombinator.com/item?id=45967814I went ahead and downloaded it, it looks to be a VSCode fork very similar to Cursor, with support for Gemini 3 Pro, Claude Sonnet 4.5, and GPT-OSS 120B models.
  • [community] https://news.ycombinator.com/item?id=48222529I much prefer using Gemini CLI in combination with vscode. It works like a charm. Now, I'll do the same with Antigravity CLI and vscode. It works fine.

Operate the product with natural-language commandsweight 2

2 (weight) × 8 (quality) × 1.0 (full) = 16.0 of 20 max

  • [claimed-docs] https://antigravity.google/product/antigravity-cliEdit, orchestrate, and build all in natural language. Tell your agents what you need, and they’ll work on getting it done.
  • [claimed-docs] https://antigravity.google/product/antigravity-ideGoogle Antigravity's Editor view offers tab autocompletion, natural language code commands, and a configurable, and context-aware configurable agent.
  • [claimed-docs] https://antigravity.google/product/antigravity-ideoffers tab autocompletion, natural language code commands, and a configurable, and context-aware configurable agent
  • [claimed-docs] https://antigravity.google/product/antigravity-2Speak your prompts. Powered by the latest Gemini Audio models, real-time transcription converts conversational speech into clearly phrased prompts.
  • [claimed-docs] https://antigravity.google/docs/artifact-reviewPlanning Mode: The agent plans thoroughly before executing tasks... produces structured implementation plans called Artifacts
  • [claimed-docs] https://antigravity.google/docs/artifact-reviewThe agent always halts and requests your explicit approval before proceeding with proposed changes.
  • [community] https://news.ycombinator.com/item?id=45967814I went ahead and downloaded it, it looks to be a VSCode fork very similar to Cursor, with support for Gemini 3 Pro, Claude Sonnet 4.5, and GPT-OSS 120B models.
  • [community] https://news.ycombinator.com/item?id=48222529I much prefer using Gemini CLI in combination with vscode. It works like a charm. Now, I'll do the same with Antigravity CLI and vscode. It works fine.

Built-in AI = 63.2 ÷ 90 × 100 = 70.2

Automation55.3/100×0.15 of the PA blend

Depth of automation primitives — rules, scheduling, bulk operations, webhooks.

Perform bulk operations across many items at onceweight 2

2 (weight) × 6 (quality) × 0.6 (partial) = 7.2 of 20 max

  • [claimed-docs] https://antigravity.google/product/antigravity-2Orchestrate multiple autonomous agents working in parallel across independent projects.
  • [claimed-docs] https://antigravity.google/product/antigravity-cliHave multiple agents working in parallel, so larger tasks get tackled faster.
  • [claimed-docs] https://antigravity.google/docs/subagentsan agent can delegate tasks—such as running tests or performing extensive codebase searches—to dedicated subagents.
  • [claimed-docs] https://antigravity.google/docs/featuresWorktree support: Projects natively support Git worktrees, allowing agents to operate in isolated background folders.
  • [claimed-docs] https://antigravity.google/docs/cli/headlessRun Antigravity CLI non-interactively to script agent tasks, integrate with CI pipelines, and capture machine-readable output.

Define rules that trigger actions automatically on eventsweight 3

3 (weight) × 7 (quality) × 1.0 (full) = 21.0 of 30 max

  • [claimed-docs] https://antigravity.google/docs/rules-workflowsAt the rule level you can define how a rule should be activated: Manual... Always On... Model Decision... Glob
  • [claimed-docs] https://antigravity.google/docs/hooksHooks allow you to run custom scripts or shell commands at specific points during Antigravity’s execution loop.
  • [claimed-docs] https://antigravity.google/product/antigravity-2Automate routine checks with Scheduled Tasks, simply define a cron schedule and the agents start and run autonomously in the background.
  • [claimed-docs] https://antigravity.google/docs/rules-workflowsRules are manually defined constraints for the Agent to follow, at both the local and global levels.

Schedule recurring jobs or workflowsweight 2

2 (weight) × 8 (quality) × 1.0 (full) = 16.0 of 20 max

  • [claimed-docs] https://antigravity.google/product/antigravity-2Automate routine checks with Scheduled Tasks, simply define a cron schedule and the agents start and run autonomously in the background.
  • [claimed-docs] https://antigravity.google/docs/featuresusers can schedule messages to be sent to their agents while they’re away

Version, review, and roll back my automationsweight 1

1 (weight) × 0 (quality) × 0.0 (none) = 0.0 of 10 max

no evidence cited — the verdict rests on absence of evidence, re-checked on refresh

Automation = 44.2 ÷ 80 × 100 = 55.3