@seclink: Compared to Mimo, Kimi is slower and more expensive ...
Summary
Kimi released the K2.7 Code model and its high-speed version, and announced API pricing. Compared to rival Mimo, it is more expensive and slower.
View Cached Full Text
Cached at: 06/26/26, 08:14 PM
Compared to Mimo, Kimi is slow and expensive… https://t.co/SewEWGfuGu
Programming Model Kimi K2.7 Code Pricing - Kimi API Open Platform
Source: https://platform.kimi.com/docs/pricing/chat-k27-code
(https://platform.kimi.com/docs/pricing/chat-k27-code#%E4%BA%A7%E5%93%81%E5%AE%9A%E4%BB%B7) Product Pricing
Here 1M = 1,000,000, the prices in the table represent the cost per 1M tokens consumed.
(https://platform.kimi.com/docs/pricing/chat-k27-code#%E6%A8%A1%E5%9E%8B%E8%AF%B4%E6%98%8E) Model Description
- Kimi K2.7 Code is Kimi’s most intelligent coding model to date, following instructions more reliably in long contexts and completing programming tasks with a higher success rate. It also supports text, image, and video input, with only thinking mode, conversation, and Agent tasks.
- Kimi K2.7 Code HighSpeed is the high-speed version of Kimi K2.7 Code. It is the same model as Kimi K2.7 Code, but with an output speed of approximately 180 Tokens/s, reaching up to 260 Tokens/s in short-context scenarios, delivering an even more extreme programming experience.
- The model has a context length of 256k, supports long thinking, and excels at deep reasoning.
- Supports automatic context caching, ToolCalls, JSON Mode, Partial Mode, and other capabilities.
Similar Articles
@seclink: Speed Comparison of coding large models: Kimi K2.7 Code HighSpeed: 180–260 t/s (Moonshot official API, balanced quality and speed, suitable for long-context coding). MiMo Ultra-High-Speed: 100…
Compares the speed performance of Kimi K2.7 Code HighSpeed (180-260 t/s) and MiMo Ultra-High-Speed (1000+ t/s) on coding tasks, pointing out that MiMo has overwhelming speed and strong quality, suitable for use with Claude Code.
Kimi K2.7 Code: 1T MoE, $0.95/M tokens, MIT license, beats Opus 4.8 on MCP tool-calling
Moonshot AI 发布了专注于编程的开放式权重模型 Kimi K2.7 Code，拥有1万亿参数和384个专家，性能在MCP工具调用上超越Opus 4.8，成本仅为十分之一。
@YRSM_Simon: This is big news! Kimi 2.6 is a generative-level model. In this age of overflowing LLM capabilities, speed will become the deciding factor in competition. Is the chip sector about to see another 'sector rotation'? 😅
Cerebras is now running Kimi K2.6, a trillion-parameter model, in enterprise trials at ~1,000 tokens/s, the fastest frontier model performance ever measured by Artificial Analysis.
@interjc: The market still needs a disruptive force; whether you use Kimi or not, the major companies' reset cycles are increasing.
Kimi releases the K3 model, featuring 2.8 trillion parameters, a million-token context window, and native multimodal capabilities. It leverages Kimi Delta Attention and Attention Residuals to enhance inference speed and training efficiency.
@jakevin7: Using Kimi K3, Maka outperforms official KimiCode by 20%. Same model, different harness — how big can the gap be? http://github.com/maka-agent/maka-agent… We ran Kimi K3 through…
On the Kimi K3 model, the open-source agent framework Maka achieves a 10% higher overall pass rate on Terminal-Bench 2.1 compared to the official Kimi Code CLI, and 20% higher on hard tasks. Through optimizations like context-budget pruning, streamlined tool surface, and concise system prompts, significant performance gains are realized. The full report and harness are open-sourced.