Poolside releases Laguna S 2.1, a 118B coding-agent model with 1M context
burny_tech · x · 2026-07-22
Poolside says Laguna S 2.1 is a 118B-A8B open-weight model with a 1M-token context window, built in-house for long-horizon agentic work.
It reports 70.2% on Terminal-Bench 2.1, ahead of several much larger models, and 59.4% on SWE-Bench Pro. The team also says it is releasing the full evaluation trajectories to support the numbers, and that real-world behavior is meant to be resourceful and persistent in coding-agent workflows.
Related event: Poolside Releases Laguna S 2.1 Open-Weight Coding Model(26 posts)→
More from coding & agent
- Building 3D Game Core Loops with Claude Fable and Three.js Agent Skills — nptacek · 2026-07-22
- Fable users demo a procedural dog simulation built with three.js agent skills — willdepue · 2026-07-22
- FDE: The $1M/Year AI Job and a 30-Day Roadmap to Master It — blaizedsouza · 2026-07-22
- Fireworks says Kimi K3 handles 72–96% of agent traffic at up to 50x lower cost — eliebakouch · 2026-07-22
- AutoIndex lifts CRUMB recall by 8.4% without changing retrievers or embeddings — mrdrozdov · 2026-07-22
- AutoIndex defines a representation program as the logic behind search indexing — mrdrozdov · 2026-07-22