MoonshotAI: Kimi Dev 72B for Software Engineering

Kimi Dev 72B excels in software bug fixing and code reasoning.

Input: text Output: text Context: 131,072 tokens Release: 2025-06-16
Discover MoonshotAI's Kimi Dev 72B, a powerful open-source language model released in June 2025. Built on Qwen2.5-72B, it excels in software engineering tasks by applying code patches and validating them through comprehensive test suites. With a generous context window of 131072 tokens, this text-to-text model is fine-tuned for precise bug fixing and code reasoning. It supports various parameters like frequency_penalty and temperature, ensuring flexible and accurate outputs for developers.

Use Cases

Here are a few ways teams apply MoonshotAI: Kimi Dev 72B in practice—from fast drafting to multimodal understanding. Adapt these ideas to your workflow.

Accurately resolve software bugs in real repositories

Enhance code reasoning capabilities

Apply and validate code patches effectively

Utilize flexible parameters for tailored outputs

Key Features

A quick look at the capabilities that make this model useful in real projects.

Open-source model fine-tuned for software engineering

Optimized using large-scale reinforcement learning

Validates code patches with full test suite execution

Achieves 60.4% on SWE-bench Verified

Supports a context window of 131072 tokens

Flexible parameters like frequency_penalty and temperature

Specs

Overview
Vendor
moonshotai
Model ID
moonshotai/kimi-dev-72b
Release
2025-06-16
Modalities & context
Input
text
Output
text
Context
131,072 tokens
Parameters & defaults

Supported parameters: frequency_penalty, include_reasoning, reasoning, response_format, structured_outputs, temperature, top_k, top_p

Defaults: temperature 0.2, top_p 0.95

Benchmark tests: MoonshotAI: Kimi Dev 72B

We ran this model against a few representative prompts to show its range. Review the outputs below and be the judge.

Text

Prompt:
Write 150 words on how AI might positively upend work, leisure and creativity
AI is poised to positively transform work, leisure, and creativity. In the workplace, automation of routine tasks will free humans to focus on complex, fulfilling roles. For example, AI chatbots handling customer service basics allows human agents to tackle more challenging issues. Data analysis by AI will enhance decision-making, boosting productivity and job satisfaction. Leisure activities will become more personalized and enjoyable. Streaming platforms already use AI for content recommendations, and AI-driven virtual assistants will manage tasks, freeing up time for hobbies and interests. Imagine a world where your leisure is tailored to your preferences, making every free moment more engaging. Creativity will flourish as AI becomes a collaborative partner. Artists can use AI to generate initial concepts or explore new styles. Writers might employ AI for editing or idea generation. The combination of human intuition and AI's data-driven insights will unlock new creative possibilities, pushing boundaries in art, design, and innovation. Overall, AI's integration into these areas will enhance efficiency, personalization, and creative potential, leading to a more fulfilling and productive society.
Run this prompt on Upend.AI

Ready to try MoonshotAI: Kimi Dev 72B?

Chat with MoonshotAI: Kimi Dev 72B
up.end
/ˌəpˈend/
verb

To “upend” means to completely disrupt, overturn, or drastically change the established order or structure of something. It implies a significant shift or alteration that can potentially have far-reaching consequences. When something is upended, it is turned upside down or transformed in a way that challenges conventional norms or expectations. The term often carries a sense of innovation, transformation, and sometimes even a hint of upheaval, indicating that the changes are not just minor adjustments but rather a fundamental reimagining of the status quo.