Moonshot has announced the release of Kimi K2, a new open-source large language model designed for high-performance coding and agentic task execution. Built using a 1 trillion parameter Mixture-of-Experts (MoE) architecture with 32 billion active parameters, the model has performed strongly across several leading open benchmarks, including SWE Bench, Tau2 and AceBench.
Positioned as one of the most capable open models for technical tasks, Kimi K2 is particularly tuned for agentic applications, where models are expected not just to respond, but to act — using tools, interpreting workflows and completing tasks with minimal instruction. This kind of task orientation is becoming more important in areas such as software engineering, automated testing and infrastructure management.
Two versions are now openly available: Kimi-K2-Base, which serves as the foundation for custom training and experimentation, and Kimi-K2-Instruct, designed for general-purpose use in chatbots or tool-enabled agent setups. While the model does not currently support multimodal input or long-form “thought-mode” processing, its performance in reasoning, code generation and structured automation has placed it at the top of several benchmarks among other open models.
Pricing for the API is tiered, with cache-hit inputs at $0.15 per million tokens and output tokens billed at $2.50 per million. Moonshot says the model is ready for integration into custom tools and workflows, allowing developers to skip complex orchestration steps. Users simply describe the task and provide the tools, and Kimi K2 handles the rest.
The release adds further momentum to the growing push for transparency and openness in AI, as more developers seek high-performance models that can be inspected, adapted and extended without platform lock-in. Kimi K2 is now live and available to use via API or open-source access.
Dear Reader,
Ledger Life is an independent platform dedicated to covering the Internet Computer (ICP) ecosystem and beyond. We focus on real stories, builder updates, project launches, and the quiet innovations that often get missed.
We’re not backed by sponsors. We rely on readers like you.
If you find value in what we publish—whether it’s deep dives into dApps, explainers on decentralised tech, or just keeping track of what’s moving in Web3—please consider making a donation. It helps us cover costs, stay consistent, and remain truly independent.
Your support goes a long way.
🧠 ICP Principal: ins6i-d53ug-zxmgh-qvum3-r3pvl-ufcvu-bdyon-ovzdy-d26k3-lgq2v-3qe
🧾 ICP Address: f8deb966878f8b83204b251d5d799e0345ea72b8e62e8cf9da8d8830e1b3b05f
🪙 BTC Wallet: bc1pp5kuez9r2atdmrp4jmu6fxersny4uhnaxyrxau4dg7365je8sy2q9zff6p
Every contribution helps keep the lights on, the stories flowing, and the crypto clutter out.
Thank you for reading, sharing, and being part of this experiment in decentralised media.
—Team Ledger Life
Community Discussion