Kimi-K2.5 via Hugging Face

Specifications

Context Window

262,144 tokens

Release Date

2026-01-01

Capabilities

AttachmentsReasoningTool callingTemperatureImage inputVideo input

Availability

Open Weights

Model Overview

Hugging Face is the leading open-source AI community and platform. Their Inference API provides easy access to thousands of models, and they maintain the most popular repository of pre-trained AI models.

Kimi-K2.5 is a kimi-family model by Hugging Face with a 262k token context window and up to 262k output tokens. It is priced at $0.6000/1M input tokens and $3.00/1M output tokens.

Key capabilities include: attachments, reasoning, tool calling, temperature, image input, video input. It supports advanced reasoning for complex multi-step tasks. It can call external tools and functions for agentic workflows.

Details

ProviderHugging Face

Model IDmoonshotai/Kimi-K2.5

Familykimi

Release Date2026-01-01

Last Updated2026-01-01

Knowledge Cutoff2025-01

Context Window262,144 tokens

Max Output262,144 tokens

Input Cost / 1M$0.6000

Output Cost / 1M$3.00

Cache Read / 1M$0.1000

More models from Hugging Face

MiniMax-M2.5

$0.3000 in · $1.20 out

GLM-5

$1.00 in · $3.20 out

Qwen3-Coder-Next

$0.2000 in · $1.50 out

Qwen3.5-397B-A17B

$0.6000 in · $3.60 out

MiniMax-M2.1

$0.3000 in · $1.20 out

GLM-4.7

$0.6000 in · $2.20 out

View all models

Frequently Asked Questions

How much does Kimi-K2.5 cost to use?

Kimi-K2.5 is priced at $0.6000/1M input tokens and $3.00/1M output tokens. Use the cost estimator on this page to calculate your expected spend based on your usage pattern.

What is a token and how does it relate to pricing?

A token is a chunk of text — roughly ¾ of a word in English. For example, "chatbot" is two tokens. LLM API pricing is based on the number of tokens you send (input) and receive (output). Input tokens include your prompts, uploaded documents, and images, while output tokens are the model's generated responses.

Why are input and output tokens priced differently?

LLM providers charge separately for input and output tokens. Output tokens are typically more expensive because generating each token requires more compute — the model must run a full forward pass for every token it produces, while input tokens are processed in parallel.

What is the context window of Kimi-K2.5?

Kimi-K2.5 supports a context window of 262,144 tokens. This is the maximum number of tokens (input + output combined) the model can process in a single request. Larger context windows let you send longer documents or maintain longer conversation histories.

How accurate is this cost estimation?

This tool provides a ballpark estimate based on per-token pricing. Actual costs may differ due to prompt caching, batched API calls, volume discounts, reasoning token overhead, and provider-specific billing rules. Use it for budgeting and comparison, not as an invoice prediction.

How does Kimi-K2.5 pricing compare to other models?

You can compare Kimi-K2.5 with other models on our LLM API pricing calculator. Use the cost estimator to see side-by-side cost breakdowns across different providers and models to find the best fit for your budget and requirements.

What factors affect my total API cost?

Your total cost depends on several factors: the number of API calls you make, the length of your prompts (input tokens), the length of generated responses (output tokens), whether you use features like image or document uploads (which add input tokens), and any provider-specific charges for caching or batch processing.

How can my team use Kimi-K2.5 via API?

You can connect your own Hugging Face API key and give your entire team access to Kimi-K2.5 through TypingMind Teams. It lets you build a unified AI workspace where team members can use Kimi-K2.5 and other models — without needing their own API keys. You stay in control of usage limits, costs, and permissions, all from a single dashboard.