Poolside Launches Laguna M.1, Free-Tier Coding Agent Model with 128K Context Window
Poolside has released Laguna M.1, its flagship coding agent model available for free on OpenRouter. The model features a 128K context window, up to 8K output tokens, and is optimized for agentic coding workflows with tool calling and reasoning capabilities.
Poolside Launches Laguna M.1, Free-Tier Coding Agent Model with 128K Context Window
Poolside has released Laguna M.1, its flagship coding agent model, available for free through OpenRouter. The model is designed for complex software engineering tasks and agentic coding workflows.
Key Specifications
Laguna M.1 offers a 128K context window (131,072 tokens) and supports up to 8K output tokens. The model has been quantized to fp8 (8-bit floating point) for efficient inference. According to Poolside, the model is optimized for tool calling and reasoning capabilities specifically tailored to software development workflows.
The free tier on OpenRouter charges $0 per million tokens for both input and output, making it accessible for developers to test and integrate into their coding workflows.
Technical Implementation
The model supports reasoning-enabled workflows, allowing it to show step-by-step thinking processes through OpenRouter's reasoning parameter. Developers can access the model's internal reasoning through the reasoning_details array in API responses.
Poolside has positioned Laguna M.1 as an agentic model, meaning it's designed to operate autonomously within coding environments with tool calling capabilities rather than simple code completion.
Release Timeline
Laguna M.1 was released on April 28, 2026, according to OpenRouter's listing. The model is currently available exclusively through OpenRouter's API, which provides normalized requests and responses across multiple providers.
What This Means
Poolside's entry into the coding agent space with a free-tier model signals increasing competition in AI-assisted software development. The 128K context window is substantial for code-related tasks, allowing developers to work with large codebases or complex multi-file operations. The fp8 quantization suggests Poolside is prioritizing inference efficiency, likely to maintain the free tier sustainably. However, without published benchmarks on standard coding evaluations like HumanEval or MBPP, it remains unclear how Laguna M.1's performance compares to established coding models from GitHub Copilot, Amazon CodeWhisperer, or Replit's offerings.
Related Articles
Nex AGI Releases Nex-N2-Pro: 397B Parameter MoE Model With 262K Context, Available Free
Nex AGI has released Nex-N2-Pro, an agentic mixture-of-experts model with 397B total parameters and 17B active parameters. The model features a 262K token context window and is available free via OpenRouter's API.
MiniMax Releases M3: 428B-Parameter Multimodal Model with 1M Context Window and 15× Decode Speedup
MiniMax has released M3, a multimodal model with approximately 428 billion parameters and 23 billion activated parameters. The model supports a 1 million token context window and uses MiniMax Sparse Attention to achieve 9× prefill and 15× decode speedups compared to its predecessor M2.
Moonshot AI releases Kimi K2.7 Code with 1T parameters, 256K context window, 30% lower thinking token usage
Moonshot AI has released Kimi K2.7 Code, a 1 trillion parameter Mixture-of-Experts model designed for long-horizon coding tasks. The model features a 256K context window and reduces thinking token usage by approximately 30% compared to its predecessor K2.6.
Apple releases AFM 3 lineup: 20B-parameter on-device model and cloud AI running on Google's Nvidia infrastructure
Apple announced five third-generation foundation models at WWDC26, headlined by AFM 3 Core Advanced—a 20-billion-parameter sparse model that runs on-device by activating only 1-4 billion parameters at a time. For the first time, Apple extended Private Cloud Compute to third-party infrastructure, with AFM 3 Cloud Pro running on Nvidia GPUs in Google Cloud.
Comments
Loading...