Google Antigravity Python SDK — v0.1.18 Release Notes

[karmel]

[0.1.18] - 2026-09-21

This release announces official Antigravity SDK local model support (LiteRT and LocalOpenAI configs are now ready to use), adds a standardized evaluation configuration preset to provide a product-agnostic default for Gemini's coding ability, introduces custom model overrides for subagents, enables schedule and background task management by default, and expands compatibility with Vertex AI service tier handling.

🌟 Key Highlights

• Antigravity SDK + Local Models Support: Official local model support is now ready to use with first-class LiteRTAgentConfig and LocalOpenAIAgentConfig configurations. Developers can execute on-device models with automated lightweight presets or integrate with local OpenAI-compatible endpoints.

  from google.antigravity import LiteRTAgentConfig, LocalOpenAIAgentConfig
  
  # On-device execution with LiteRT
  litert_config = LiteRTAgentConfig(model_path="/path/to/gemma-4-26b.bin")
  
  # Local OpenAI-compatible server
  local_config = LocalOpenAIAgentConfig(base_url="http://localhost:8000/v1")

• Standardized Evaluation Preset (AgentConfig.eval()): Adds a standardized, benchmark-ready preset that configures autonomous permissions, benchmark retry behavior, daemon execution in commands, and strips image generation and subagents to focus on core coding evaluations.

  from google.antigravity import LocalAgentConfig
  
  config = LocalAgentConfig().eval()

• Custom Subagent Models: Allows developers to assign specific model targets to subagents independently of the root agent configuration.

  from google.antigravity import SubagentConfig
  
  researcher = SubagentConfig(name="researcher", model="gemini-2.5-pro")

• Built-in Schedule Tool: Enables the schedule tool by default across standard tool groups, permitting agents and subagents to set timers and cron-based background jobs alongside task management.

  from google.antigravity import BuiltinTools, LocalAgentConfig
  
  # BuiltinTools.SCHEDULE is enabled by default; exclude if not desired
  config = LocalAgentConfig(disabled_tools=[BuiltinTools.SCHEDULE])

• Automatic Lightweight Presets for LiteRT: Instantiating LiteRTAgentConfig now automatically applies optimized lightweight presets—including reduced prompt overhead and synchronous context compaction—without requiring an explicit call to .lightweight().

  from google.antigravity import LiteRTAgentConfig
  
  # Automatically configures lightweight presets for local execution
  config = LiteRTAgentConfig(model_path="/path/to/gemma-4-26b.bin")

---

📋 Detailed Changes

Features & Enhancements

• Task Management Pairing: Automatically enables the manage_task tool whenever run_command or schedule is active, allowing background task lifecycle management.
• LiteRT & Local Model Examples: Added getting-started guides and end-to-end examples demonstrating on-device execution with LiteRTAgentConfig (Gemma 4 26B) and OpenAI-compatible local endpoints via LocalOpenAIAgentConfig.
• Sandbox Availability Warning: Added a session startup warning when enable_sandbox=True is requested on an environment or OS backend where sandbox isolation cannot be enforced.
• Extended JSON Schema Normalization: Added schema normalization support for OpenAPI and JSON Schema Draft 7 / 2020-12 keywords (such as multipleOf, prefixItems, and dependentSchemas) and prevented accidental mutation of uppercase sample values.
• Optimized ToolRunner Coercion: Improved type resolution for closure-scoped tools and forward-referenced ToolContext parameters, and introduced TypeAdapter caching to accelerate tool call execution.
• Top-Level ServiceTier Export: Re-exported ServiceTier at the root package namespace for easier import parity.

Model & Default Changes

• Excluded ASK_QUESTION from Default Tools: Excluded BuiltinTools.ASK_QUESTION from BuiltinTools.default(). Default configurations run autonomously; agents in headless workflows will no longer attempt interactive user prompts. To re-enable interactive questions, explicitly pass BuiltinTools.ASK_QUESTION in enabled_tools.
• Single Compaction Threshold Dial: Simplified CompactionConfig to a single token_threshold property, deprecating legacy context token limits and interval dials. To configure context compaction, specify CompactionConfig(token_threshold=...).
• Removed Deprecated modified_arguments_json: Removed the legacy JSON string fallback in tool hook interception in favor of modified_args.

Bug Fixes

• Service Tier Ingestion: Fixed an unhandled ValueError when connecting via gateways reporting unlisted backend service tiers (such as Vertex AI PROVISIONED_THROUGHPUT) by safely ignoring unknown tier values while preserving token counts.