Kimi K2.6 · Moonshot AI

What does Kimi K2.6 cost to run in a coding agent?

Kimi K2.6 costs about $0.024 per average coding-agent request on our measured traffic, at OpenRouter's listed prices of $0.65 per million input and $3.41 per million output tokens. It scores 27 on the Artificial Analysis Intelligence Index, 32nd of the 48 models we track. Source: Artificial Analysis (artificialanalysis.ai)

List prices: OpenRouter listed price, checked 2026-10-13. Request costs: Standard Compute gateway data, 79,301 coding-agent requests from 75 users, Sep 13 to Oct 12, 2026. How we compute this

Run current Claude, GPT, Gemini and open models in your agent

Or let the router send each step to the cheapest model that can do it, on providers we vet for price, uptime and speed. Plans from $19 a month.

See plans

What Kimi K2.6 costs per coding-agent request

Cost at list price	Kimi K2.6
Prompt under 100K tokens50% of requests, 87% of input cached	$0.014
Prompt of 100K to 272K39% of requests, 88% of input cached	$0.037
Prompt over 272K tokens11% of requests, 86% of input cached	n/a
Average request	$0.024
100 requests	$2.41
Median session16 requests; a session ends after 10 idle minutes	$0.28
A month of regular usemedian of 20 users active on 5+ days	$30.86

Each logged request priced at the listed rates above for its own token counts, including cache reads and any higher rate for long prompts. Average output: 698 tokens per request. 12% of requests were larger than Kimi K2.6's 262K context window and are left out of its figures.

Kimi K2.6 benchmarks, speed and latency

Artificial Analysis	Value	Rank
Intelligence Index10 evaluations, 0 to 100	27	32 of 48
Terminal-Bench 4.0agentic coding in a terminal	0.5%	33 of 39
Output speedmedian	43 tokens/s	41 of 44
Time to first answer tokenincludes thinking time	105 s	31 of 44

Source: Artificial Analysis (artificialanalysis.ai). Measured variant: Kimi K2.6 (Reasoning), speed on Kimi. Fetched October 13, 2026. Rank is among the models on this site, older versions included. Artificial Analysis runs Terminal-Bench 4.0 in its own setup, so its scores differ from the tbench.ai leaderboard.

What to watch for

Its 262K-token window is smaller than 12% of our coding-agent requests, which are left out of the figures above, so long agent sessions will hit the limit. The price is OpenRouter's listed rate on October 13, 2026, which can be a third-party host's; Moonshot AI's own API may charge differently. It is weak on long agentic coding tasks (see Terminal-Bench 4.0 above), so a cheap request can turn into more steps and retries.

About Kimi K2.6

Kimi K2.6 is Moonshot AI's open-weight general model from April 20, 2026, for reasoning, writing and agentic work, with a 262K context. It scores 27 on the Artificial Analysis Intelligence Index and 0.5% in Artificial Analysis's Terminal-Bench 4.0 run, at about 43 output tokens per second, and Artificial Analysis lists it as superseded. At OpenRouter's listed price it costs $0.024 per request in our measurements, and its 262K context is smaller than 11.7% of the coding-agent requests we measured.

Sources: Artificial Analysis

Kimi K2.6 FAQ

How much does Kimi K2.6 cost per coding-agent request?

$0.024 on average at OpenRouter's listed price, $2.41 per 100 requests, and $0.28 for a median 16-request session. That is 79,301 real coding-agent requests (Sep 13 to Oct 12, 2026) priced at OpenRouter's listed rates; the median request sent 99,534 input tokens, 87% of input was read from cache, and requests returned 698 output tokens on average.

Does Kimi K2.6 charge more for long prompts?

No. Kimi K2.6 bills the same rate per token up to its 262K-token context window, which is smaller than 12% of the coding-agent requests we measured.

How fast is Kimi K2.6?

Artificial Analysis measures Kimi K2.6 (Reasoning) at 43 output tokens per second, with 105 seconds to the first answer token, thinking time included, on Kimi's API. Source: Artificial Analysis (artificialanalysis.ai).

Is Kimi K2.6 good for AI agents?

General writing and reasoning on an open-weight model. On our scoring, which is calibrated to public benchmarks, it rates 4.6/10 for agentic ability and 6.0/10 for output quality. Open weights.

What is Kimi K2.6's context window?

Kimi K2.6 has a 262K-token context window. Moonshot AI is the maker.

Is Kimi K2.6 open source?

Yes. Kimi K2.6 ships with open weights, so you can self-host it or run it through a hosted provider such as OpenRouter. Check the license before commercial use.

What are the downsides of Kimi K2.6?

The main trade-offs: Weak on long agentic tasks (0.5% in Artificial Analysis's Terminal-Bench 4.0 run); Slow (about 43 output tokens/s, per Artificial Analysis); 262K context, smaller than 11.7% of the coding-agent requests we measured. It's strongest for general writing and reasoning on an open-weight model.