Researchers escaped the sandboxes of Cursor, Codex, Gemini CLI and Antigravity without breaking them. Three vendors patched; Google downgraded its two.
Poolside's 118B-parameter Laguna S 2.1 matches larger models on coding benchmarks, runs on a single desktop, and targets enterprises wanting to self-host.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results