Kimi K2.6
MoonshotAiFlagship
ThinkingTool UseVisionStructured Output
About this model
Kimi K2.6 is Moonshot AI's next-generation multimodal model, designed for long-horizon coding, coding-driven UI/UX generation, and multi-agent orchestration. It handles complex end-to-end coding tasks across Python, Rust, and Go, and...
Performance Tier
Flagship
Kimi K2.6 is a flagship model from MoonshotAi : the most capable in their lineup.
Best-in-class model from this provider. Highest performance across benchmarks, ideal for demanding tasks.
Pricing
This model is included in Elosia plansModerate
Moderate cost. A balanced choice for regular use without constant cap watching.
| Type | per 1M tokens |
|---|---|
| Input (prompt) | $0.750 |
| Output (completion) | $3.50 |
| Cache read | $0.150 |
Capabilities
Context Length262K
Max Output Tokens16K
TokenizerOther
Inputtext, image
Outputtext
Release DateApril 20, 2026
Benchmarks
General Intelligence
MMLU
Not reported
GPQA Diamond
90.5%
Mathematics
MATH-500
Not reported
AIME 2026
96.4%
Programming
HumanEval
Not reported
SWE-bench Verified
80.2%
SWE-bench Multilingual
76.7%
Reasoning
IFEval
Not reported
Humanity's Last Exam
54%
Multimodal
MMMU-Pro
79.4%
Agentic
SWE-bench Pro
58.6%
Terminal-Bench 2.0
66.7%
Recommended Use Cases
CodingAnalysisResearchGeneral Chat
Strengths
- Top-tier agentic coding — SWE-bench Verified 80.2%
- Open-weight 1T MoE (32B active) with 256K context window
- Excellent graduate-level reasoning (GPQA Diamond 90.5%)
- Strong long-horizon task orchestration via agent swarm
Limitations
- Moonshot official evaluation focused on agentic and reasoning benchmarks — MMLU and MATH-500 are not part of the reported suite
- Smaller Western ecosystem than Claude or GPT families
- High output pricing for a Moonshot model ($3.50/M output)
Resources
This model may use your data for training
