Why This Matters
Laguna S 2.1 represents a major advancement in AI model development, offering enhanced reasoning and long-horizon capabilities with a massive context window of up to 1 million tokens. Its rapid deployment and competitive performance on coding benchmarks highlight its potential to transform AI applications in complex problem-solving and extended reasoning tasks, impacting both the tech industry and end-users. This development underscores the ongoing trend toward more powerful, efficient, and context-aware AI models that can better understand and generate sophisticated outputs.
Key Takeaways
- Laguna S 2.1 features a 118B parameter MoE architecture with a 1 million token context window.
- It was developed and launched in under nine weeks, demonstrating rapid deployment capabilities.
- The model performs competitively on long-horizon coding benchmarks, rivaling larger models.
Today we’re releasing Laguna S 2.1, a significant step forward in our development of models that pursue longer horizon work and make effective use of reasoning.
Laguna S 2.1 is a 118B total parameter Mixture-of-Experts (MoE) model with 8B activated parameters per token and supports a context window of up to 1M tokens in thinking and no-thinking modes. It went from the start of training to launch in under nine weeks, and on long-horizon coding benchmarks it holds its own against models many times its size. For every benchmark score we publish today, we are releasing full trajectories for every trial in the final evaluation set at trajectories.poolside.ai .
Laguna S 2.1 118B-A8B
Tencent Hy3 295B-A21B
Inkling 975B-A41B
Nemotron 3 Ultra 550B-A55B
DeepSeek-V4-Pro Max 1.6T-A49B
Kimi K3 2.8T-A50B
Qwen 3.7 Max —
Muse Spark 1.1 —
... continue reading
GoKawiil