Question 1

What is Kimi K2.5?

Accepted Answer

Kimi K2.5 is a model from Moonshot, released on Jan 27, 2026. Kimi K2.5 is Moonshot AI's native multimodal model, delivering state-of-the-art visual coding capability and a self-directed agent swarm paradigm. Built on Kimi K2 with continued pretraining over approximately 15T mixed visual and text tokens, it delivers strong performance in general reasoning, visual coding, and agentic tool-calling.

Question 2

How much does Kimi K2.5 cost?

Accepted Answer

On the Respan gateway, input starts at $0.45 per 1M tokens on OpenRouter, and output starts at $2.25 per 1M tokens on OpenRouter, after gateway discounts. Averaged across 7 providers, it costs $0.60 per 1M input tokens and $2.97 per 1M output tokens.

Question 3

Which providers serve Kimi K2.5 on Respan?

Accepted Answer

7 providers serve Kimi K2.5: OpenRouter, TogetherAI, IO.net, Novita, Tera, Featherless, and 1 more. TogetherAI and Tera require your own provider API key (BYOK).

Question 4

What is Kimi K2.5's context window?

Accepted Answer

Kimi K2.5 accepts up to 262K tokens of context on its largest provider route.

Question 5

How do I use Kimi K2.5 with the Respan gateway?

Accepted Answer

Point any OpenAI-compatible SDK at https://api.respan.ai/api/ with your Respan API key and set the model to "kimi-k2.5". To set it up with a coding agent, copy the agent setup prompt at the top of this page.

Question 6

How fast is Kimi K2.5?

Accepted Answer

Over the last 30 days of requests through the Respan gateway, the fastest provider was IO.net at 27 tokens per second, and the lowest response latency was 1.4 seconds on Novita. Speed varies with request size and provider.

Question 7

Where does Kimi K2.5 rank on the Respan Index?

Accepted Answer

Kimi K2.5 scores 27.1 out of 100 on the Respan Index v1.1, ranking #73 of 138 models. By category: Coding 9.9, Agents 32.0, Reasoning 22.5, Knowledge & language 33.9, and Human preference 51.0. The Respan Index combines independent and lab-reported benchmark results into one score, using benchmark data collected by llmmetric.com.

Question 8

How does Kimi K2.5 score on benchmarks?

Accepted Answer

Kimi K2.5 has 26 published benchmark scores, including LiveCodeBench (v6) 85%, OSWorld-Verified 63.3%, and GPQA-Diamond 87.6%. Each score links to where it was published; the data is collected by llmmetric.com.

Provider	Context	Input	Output
OpenRouter`openrouter/moonshotai/kimi-k2.5`	262K	$0.45/M	$2.25/M
TogetherAIBYOK only`together_ai/moonshotai/Kimi-K2.5-fp4`	262K	$0.50/M	$2.80/M
IO.net`ionet/moonshotai/Kimi-K2.5`	262K	$0.52/M	$2.69/M
NovitaSave 10%`novita/moonshotai/kimi-k2.5`	262K	$0.54/M	$2.70/M
TeraBYOK only`tera/moonshotai/Kimi-K2.5`	262K	$0.60/M	$3.00/M
Featherless`featherless_ai/moonshotai/Kimi-K2.5`	262K	$0.80/M	$3.40/M
Inference.net`inference_net/kimi-k2.5`	262K	$0.80/M	$3.95/M

Kimi K2.5

Providers

Certified where it counts.

Pricing

Cost by volume

Throughput and latency

Output speed

Response latency

Benchmarks

Respan Indexv1.1

Coding8

Apps

Kimi K2.5 FAQ

Works with your entire stack

Built for AI agents.
Break less.
Ship more.

Agents4

Reasoning4

Math2

Knowledge2

Multimodal2

Arena3

Other1