Chinese artificial intelligence company Moonshot AI has introduced Kimi K3.1 Pro, a new open-weight AI model designed to compete with leading proprietary systems in coding, agentic workflows and knowledge-based tasks. The company claims that its latest model delivers competitive performance while offering lower usage costs than its predecessor.
According to benchmarks published by Moonshot AI, Kimi K3.1 Pro achieved a score of 65.7% on Terminal-Bench 4.0, a benchmark focused on evaluating AI models’ ability to complete tasks in terminal environments. The reported result places it ahead of OpenAI’s GPT-6 Astra, which scored 58.2%, and just behind Anthropic’s Claude Opus 5.5 at 66.4%.
These results position Kimi K3.1 Pro as a notable contender in the increasingly competitive AI market. However, the figures come from the company’s own reported evaluations, and independent testing would be needed to establish how consistently the model performs across different tasks and real-world applications.
Moonshot AI described K3.1 Pro as the first model in its new K3.1 series and its most capable 3-trillion-parameter-class open-weight model to date. The release represents the company’s continued effort to develop advanced AI systems that can compete with established providers on performance, flexibility and cost.
One of the model’s major features is its reported one-million-token context window. A context window of this size allows an AI system to process large amounts of information within a single interaction, potentially making it useful for analysing lengthy documents, reviewing extensive codebases and handling complex research projects.
The company has also highlighted improvements in coding, 3D creation, financial research and agentic tasks. These capabilities are particularly relevant as AI systems move beyond answering individual questions towards completing multistep assignments with less direct human intervention.
In software development, stronger coding performance can help developers analyse code, identify problems and work through complex programming tasks. In financial research, AI tools can assist with processing large datasets and synthesising information, although their outputs still require appropriate verification.
Agentic capabilities are another important area of competition among AI developers. Such systems are designed to plan and execute sequences of actions, interact with tools and work towards defined objectives. Their effectiveness depends not only on benchmark scores but also on reliability, accuracy and the ability to recover from errors.
Moonshot AI has also introduced Kimi K3.1 Pro Highspeed, a faster version of the model. The company says the Highspeed variant operates two to three times faster than the standard version, potentially making it more suitable for users who prioritise response speed and workflow efficiency.
Both versions are available through Kimi.ai, Kimi Work, Kimi Code and the Kimi API, according to the announcement. The company has also stated that the full model weights are scheduled for release by November 4, allowing developers to anticipate broader access to the model’s underlying parameters.
The open-weight approach could make Kimi K3.1 Pro attractive to developers and organisations seeking greater control over how AI models are deployed. Depending on the eventual licence and hardware requirements, open weights can provide opportunities for customisation, research and self-hosted applications.
Cost is another important consideration. Lower API prices can make advanced AI capabilities more accessible to startups, developers and businesses operating under tight budgets. However, the overall cost of using a model also depends on token consumption, hardware requirements, speed and the amount of human review needed.
The reported Terminal-Bench results suggest that competition between AI companies is increasingly focused on practical task completion rather than conversational ability alone. Still, a single benchmark cannot establish that one model is universally better than another, particularly across different workloads and evaluation methods.
Kimi K3.1 Pro’s reception will ultimately depend on independent performance testing, pricing details, reliability and real-world developer experience. Its launch nevertheless highlights the growing competition to deliver powerful AI systems that combine advanced capabilities with more affordable access.