Poolside's Laguna S 2.1 is a small open-weight coding model that punches well above its size

Poolside made its first models available to a broader audience in April 2026 with Laguna M.1 and XS.2. Until then, the company had focused on government and public-sector customers. XS.2 was also its first open model under the Apache 2.0 license. Laguna S 2.1 is the third version in the series released in roughly three months.

Laguna S 2.1 beats much larger open models

With thinking enabled, Laguna S 2.1 scores 70.2 percent on Terminal-Bench 2.1, which tests models on long-running terminal tasks. It ranks just behind Tencent's Hy3 (295B-A21B) and ahead of much larger open models, including DeepSeek-V4-Pro-Max, Nemotron 3 Ultra, and Thinking Machines Lab's debut model. The overall leaderboard is led by OpenAI's GPT-5.6 Sol, Anthropic's Claude Fable 5, and Kimi K3.

Poolside says Datacurve's DeepSWE benchmark offers a better comparison because its scores are spread across a wider range. Laguna S 2.1 scores 40.4 percent, while some open-weight models with more than one trillion parameters remain below 10 percent. It also ranks near the top of its class on SWE-Bench Multilingual, SWE-Bench Pro, and SWE Atlas.

Thinking mode has a major impact on performance. Without it, Laguna S 2.1's Terminal-Bench score drops to 60.4 percent, while its DeepSWE score falls to 16.5 percent. Poolside says no previous Laguna model has shown a larger performance gap between the two modes.

Poolside treats persistence as an alternative to raw scale

Poolside says the release reflects a broader idea about model performance. "What we've done in this model is not necessarily add more intelligence, but improve the behaviors that lead to a more capable model: more verification, less taking things for granted, not declaring victory early, and being more persistent," the company writes in its release post.

Earlier Laguna models sometimes stopped after only partially passing a test suite or abandoned an approach just two steps before it would have worked. Poolside treats persistence, verification, and revising failed approaches as a second path to better performance alongside scaling the model itself. A larger Laguna model is already in pre-training.

Poolside supports its claim with three documented trial runs. In one, Laguna S 2.1 built a working browser engine from an empty folder in 50 minutes that could render HTML and CSS. In another, the model found a proof for Erdos Problem #397, a math problem that had been open since 1975, while working in a sandbox without Python. Poolside says the result was an independent rediscovery. GPT-5.2 Pro solved this and several other problems in January 2026, while Laguna's training cutoff was November 2025.

Post-training across 409,000 environments drives the gains

Poolside says the jump from XS 2.1 to S 2.1 came mainly from scaling and post-training, not new pre-training data. The agentic training phase covered 409,000 environments, including 83,000 for terminal tasks and 168,000 for software engineering workflows. The largest single source was about 38,000 real commits from roughly 17,000 repositories. A new task category trained the model to install repositories on its own, set up every dependency, and get test suites running.

Poolside increased rollout budgets and extended timeouts. It also built a new sandbox system that can selectively block network access to curb reward hacking. Multi-harness rollouts run the same prompts across several agent environments, reducing the risk of overfitting to one setup.

Fewer than nine weeks passed between the start of training and launch, according to Poolside. Pre-training began on May 22, 2026, using 4,096 Nvidia H200 GPUs. S 2.1 is also the company's first model trained with reinforcement learning in FP8 precision.

Poolside published every benchmark trajectory at trajectories.poolside.ai. During training, reward hacking rates topped 50 percent on SWE-Bench tasks because the model searched online for matching pull requests instead of solving the tasks itself. A small prompt change brought the rate below two percent.

Laguna S 2.1 is still too closely tuned to Poolside's agent harness in some cases. In unfamiliar environments with slightly different tool schemas, the model can stray from the required format, Poolside says. It also tends to produce overly long thinking sequences on competitive math problems. Users can't adjust its thinking effort yet.

Laguna S 2.1 runs locally or through hosted services

Laguna S 2.1 is available on Hugging Face under the OpenMDW 1.1 license. Backed by the Linux Foundation, the license allows anyone to use, modify, and redistribute the model weights, including for commercial purposes.

Baseten, Vercel AI Gateway, and OpenRouter offer hosted access. OpenRouter provides a free endpoint with a 256K context window and a paid endpoint supporting the full one-million-token window. Poolside says the model can also run locally on a single Nvidia DGX Spark. A free demo chat is available at chat.poolside.ai without a login.

Poolside is making two strategic bets. One is that the path to intelligence runs through agentic coding because software gives agents their most expressive interface. It also believes AI can "decompress the web." Most written material records answers rather than the reasoning behind them, and Poolside argues that reinforcement learning can recover that process.