Sulat.com
AI models
$10 off the fastest DeepSeek V4.1 Flash, Kimi K3 and GLM 5.3 from Synthetic
SiliconFlow (China) logo

Model details

Pro/moonshotai/Kimi-K2.6

Kimi K2.6 is the latest release in Moonshot AI's open-source Kimi series, launched in late April 2026, with independent reporting framing the model as a push into stronger coding performance and multi-agent workflows. Coverage from Yicai Global describes the launch as extending Moonshot's line of open-source large models and positions K2.6 around agent-style tasks where the model orchestrates tools and chained reasoning steps. The OpenRouter listing confirms a roughly 262,000-token context window, making the model suitable for long-document analysis, large codebases, and multi-turn agent sessions that need to retain substantial state across steps.

On aggregator routes, Kimi K2.6 is priced at roughly $0.56 per million input tokens and $2.36 per million output tokens, a notably lower cost than flagship peers like GPT-5.6 Terra Pro that share the same routing infrastructure, which makes it a practical choice for high-volume coding and agent pipelines where per-token spend matters. The combination of a generous context window, coding-focused positioning, and agent-friendly capabilities such as tool calling and structured output gives K2.6 a clear fit for developer-assistant applications, automated refactoring, and orchestrated research workflows that need sustained reasoning over large inputs. As an open-source release from Moonshot AI, it also offers flexibility for teams that want to self-host or fine-tune for domain-specific agent behavior.

SiliconFlow (China)Pro/moonshotai/Kimi-K2.6kimi

Quick Info

Powered by
Provider
SiliconFlow (China)
Model key
Pro/moonshotai/Kimi-K2.6
Release date
Apr 21, 2026
Last updated
Apr 21, 2026
Input modalities
Output modalities
Capabilities

Cost

Input token cost
$0.95
Output token cost
$4.00

Limits

Output tokens
262,000 tokens
Context window
262,000 tokens

Transparent token rates

Compare Pro/moonshotai/Kimi-K2.6 pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about Pro/moonshotai/Kimi-K2.6

SiliconFlow (China)

Coverage

The Interconnects "Latest open artifacts (#21)" newsletter (May 16, 2026) includes Kimi K2.6 in its headline listing of an open-model release wave alongside Gemma 4, DeepSeek V4, MiMo 2.5, and GLM-5.1, framing the month as an "open model bonanza." The substantive technical content in the excerpt focuses on CAISI's eval The piece does not surface Kimi K2.6-specific benchmark numbers in the supplied excerpt — analysis centers on DeepSeek V4 and aggregate open-vs-closed gap — so its value here is as ecosystem context confirming K2.6 as part of the May 2026 open release cohort and the surrounding third-party evaluation landscape, rather

SiliconFlow (China)

CoverageRelease Notes

(Yicai) April 21 -- Chinese artificial intelligence startup Moonshot AI launched Kimi K2.6, the latest addition to its Kimi series of open-source large...

Videos about Pro/moonshotai/Kimi-K2.6

More models around Pro/moonshotai/Kimi-K2.6