Laguna XS.2

At a glance

What Laguna XS.2 does

Tool use, Streaming, Structured output, Agentic coding, Native reasoning, Preserved thinking, SWA + global attention, FP8 KV cache, Ollama/MLX support, Self-hostable

Best for

Local agentic coding on Mac/laptop (runs on 36GB), SWE-bench tasks, long-horizon autonomous coding, Zed/JetBrains integration via ACP

Works well with

Conflicts & caveats

• Trae's built-in model menu may not include Laguna XS.2 — Trae primarily ships with Claude/GPT/Gemini/DeepSeek.
• Google Antigravity is built around Gemini 3 Pro; Claude and recent GPT-5 models also work, but Laguna XS.2 is not in the supported model list.
• Kiro's spec mode is tuned for Claude models — Laguna XS.2 may work but spec-driven flows are best with Claude.
• Cursor's built-in model menu may not include Laguna XS.2 — verify it's available in your Cursor model picker.
• OpenAI Codex (VS Code) is tied to OpenAI's GPT-5.1 Codex family — Laguna XS.2 is not selectable inside this integration.
• Gemini CLI is locked to Google Gemini models — your LLM is Laguna XS.2.
• Privacy conflict: Self-hosted Laguna XS.2 sends code to cloud @codebase (Cursor). Use local context (Continue indexing, ChromaDB, LanceDB, pgvector, Vespa self-hosted) for true privacy.
• Claude Code only works with Claude models — your LLM is Laguna XS.2.
• Devin uses Cognition's proprietary model internally — your selected Laguna XS.2 will be ignored.
• OpenAI Codex (Cloud) runs GPT-5.1 Codex internally — your selected Laguna XS.2 will be ignored.
• Jules runs Gemini models internally — your selected Laguna XS.2 will be ignored.
• The Claude Code SDK is tuned for Claude models — Laguna XS.2 may work via the model adapter but loses Claude-specific features (thinking, prompt caching, sub-agents).
• Lovable uses a fixed internal model — your selected Laguna XS.2 will be ignored.
• Replit Agent 3 uses its own model selection internally — your selected Laguna XS.2 will be ignored.