kimi/kimi-k2.5

kimi/kimi-k2.5
VisionToolsJSONReasoning
by kimi

Moonshot Kimi K2 (0905 baseline) — 1T-param MoE chat model with 32B active per pass, 256k context, balanced performance.

ctx262.1K tokens
Max output32.8K
Inputtext + image
Outputtext
p50 TTFT3.71 s
INPUT$0.60/ 1M tokens
OUTPUT$3.00/ 1M tokens
p50 TTFT3.71 s7d
p95 TTFT10.00 s7d
TRAFFIC6.8Mtokens / 7d

Kimi K2.5 is a multimodal language model created by the provider Kimi. It accepts both text and image inputs and is designed to handle long-context tasks with a context window of 262,144 tokens. The…

What is Kimi K2.5 and who developed it?

What are the key specifications of Kimi K2.5?

Who is the target audience for Kimi K2.5?

How does the image input work with Kimi K2.5?

Code samples

Call from any SDK

OpenAI-compatible — keep the SDK you already use

  • OpenAI SDKhttps://api.orcarouter.ai/v1
from openai import OpenAI

client = OpenAI(
    base_url="https://api.orcarouter.ai/v1",
    api_key="$ORCAROUTER_API_KEY",
)

response = client.chat.completions.create(
    model="kimi/kimi-k2.5",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Supported parameters

  • frequency_penalty
  • include_reasoning
  • max_completion_tokens
  • max_tokens
  • n
  • presence_penalty
  • prompt_cache_key
  • reasoning
  • response_format
  • safety_identifier
  • stop
  • stream
  • stream_options
  • temperature
  • thinking
  • tools
  • top_p

Pricing

Input / 1M tokens$0.600
Output / 1M tokens$3.00
Cache read / 1M$0.100
CurrencyUSD

Cost calculator

Tokens / month10MM
Input share70%%
Estimated / month $13.20 · With prompt caching $11.45

Estimate based on list price

Token & cost estimator

Input tokens: 20Cost per request: $0.001512

Estimate only — actual token counts depend on the provider's tokenizer.

Performance

p50 TTFT
3.71 s
Output speed
61.7 tok/s
p95 TTFT
10.00 s
Error rate
0%

Public benchmarks

39.5
AA Coding
Better than 49% of models compared
#54 of 106
46.8
AA Intelligence
Better than 56% of models compared
#48 of 110
GPQA Diamond
87.9
Humanity's Last Exam
29.4
IFBench
70.2
Long-Context Recall
65.3
SciCode
49.0
TerminalBench Hard
34.8
τ²-Bench
95.9
Source: artificialanalysis.ai

How it compares

kimi/kimi-k2.5Kimi K3kimi/kimi-k2.6Kimi K2.7 Code
Input $/M$0.60$3.00$0.95$0.95
Output $/M$3.00$15.00$4.00$4.00
Context262K1.0M262K262K
Quality7/109/108/108/10
Compare side-by-sideCompare side-by-sideCompare side-by-sideCompare side-by-side

FAQ

What is the pricing for Kimi K2.5 on OrcaRouter?
Kimi K2.5 is priced at $0.60 per 1 million input tokens and $3.00 per 1 million output tokens. These are the provider rates with zero markup added by OrcaRouter. There are no additional fees or tiered pricing.
What is the context window size of Kimi K2.5?
The context window is 262,144 tokens, including both text and image tokens. The maximum output length is 32,768 tokens per request.
What are the strengths of Kimi K2.5?
The model has a large context window, supports both text and image inputs, and achieved a strong score of 95.9 on the τ²-Bench benchmark, which measures tool-use reasoning. It is particularly suited for long-document analysis, multimodal tasks, and agentic workflows.
How does Kimi K2.5 compare to other models?
With a context window of 262K tokens, it is competitive with other long-context models. Its τ²-Bench score suggests strong performance in tool-use reasoning. However, no other benchmark numbers are available, so comparison on other dimensions (e.g., general knowledge, coding) is not possible from the provided facts.
Does OrcaRouter cache or store my data when using Kimi K2.5?
The provided facts do not include details about data handling. Please refer to OrcaRouter's privacy policy and terms of service for information on data storage, retention, and caching practices.
How do I access Kimi K2.5 via an OpenAI-compatible API?
Set the base URL to https://api.orcarouter.ai/v1, use your OrcaRouter API key, and specify the model ID 'kimi/kimi-k2.5' in your requests. The endpoint supports standard OpenAI chat completions format, including streaming and image input via image_url parts.
Can I use Kimi K2.5 for free?
No, Kimi K2.5 is a paid model. There is no free tier. You must have credits or a billing account with OrcaRouter to use it.
What image formats are supported by Kimi K2.5?
The model accepts images as part of the input. While specific formats are not listed, typical implementations support common formats like JPEG, PNG, GIF, and WEBP. Images are usually provided as URLs or base64-encoded data in the user message content.

Embed this badge

kimi/kimi-k2.5$0.60/M in3705ms p50via OrcaRouter
HTML <a href="https://www.orcarouter.ai/models/kimi/kimi-k2.5" target="_blank"> <img src="https://www.orcarouter.ai/embed/kimi/kimi-k2.5.svg" alt="kimi/kimi-k2.5 on OrcaRouter" /> </a>
Markdown [![kimi/kimi-k2.5](https://www.orcarouter.ai/embed/kimi/kimi-k2.5.svg)](https://www.orcarouter.ai/models/kimi/kimi-k2.5)

Model card as data

GET /api/public/models/kimi/kimi-k2.5Open
Machine-readable:/llms.txt/llms-full.txt