Kimi K2.6 Thinking is the reasoning-optimized K2.6 variant for deeper multi-step planning and execution. It is tuned for long-horizon coding and design workflows, including complex orchestration across many specialized sub-agents and autonomous end-to-end output generation.
Added Apr 16, 2026
Model weightsContext Window
256.0K
Max Output
65.5K
Avg output tokens (7d)
3.7K tokens
Input Price (Auto)
$0.59/1M
Output Price (Auto)
$2.47/1M
Cache Read (Auto)
$0.099/1M
Capabilities
Benchmarks
Benchmarks
Performance metrics and benchmarks
Sourced from Artificial Analysis.
Intelligence Index
31.3
Coding Index
61.8
Agentic Index
22.1
Agentic work
AutomationBench-AA
Workflow automation with guardrail penalties
13.0%
Better than 41% of models compared
AA-Briefcase
Agentic knowledge work (Elo)
820 Elo
Better than 42% of models compared
GDPval-AA v2
Economically valuable tasks (Elo)
1115 Elo
Better than 58% of models compared
Document reasoning
GDP.pdf
Professional PDF reasoning: all-pass rate
13.0%
Better than 59% of models compared
AA-LCR v1.1
Long context reasoning with updated grading
81.0%
Better than 93% of models compared
Reasoning
HLE
Humanity's Last Exam
37.5%
Better than 89% of models compared
IFBench
Instruction-following benchmark
76.0%
Better than 95% of models compared
CritPt
Research-level physics reasoning
8.0%
Coding
Terminal-Bench Hard (legacy)
Legacy fallback · Agentic coding and terminal use
43.9%
Better than 91% of models compared
SciCode
Python programming for scientific computing
51.5%
Better than 64% of models compared
Knowledge
AA-Omniscience Accuracy
Proportion of correctly answered questions
32.6%
AA-Omniscience Hallucination Rate
Rate of incorrect answers among non-correct responses
40.5%
Legacy benchmarks
GPQA Diamond (legacy)
Graduate-level scientific reasoning
91.1%
Better than 92% of models compared
T²-Bench Telecom (legacy)
Conversational AI agents in dual-control scenarios
95.9%
Better than 96% of models compared
AA-LCR (unversioned / legacy)
Long context reasoning evaluation
81.0%
Better than 93% of models compared
GDPval-AA (unversioned / legacy)
Economically valuable tasks
30.8%
Last updated Sep 9, 2026
Artificial AnalysisProviders
Choose explicit providers for this model. Auto routing remains available as the default option.
Loading provider options…
Related text models
Compare Kimi K2.6 Thinking with similar models from the same provider or model family.
Kimi K3
moonshotai/kimi-k3Kimi K3 is Moonshot AI's open-weight, always-thinking multimodal model for long-context reasoning, coding, tool use, and native image/video understanding.
Kimi K2.7 Code High-Speed
moonshotai/kimi-k2.7-code-highspeedKimi K2.7 Code High-Speed is the accelerated coding-focused variant tuned for fast agentic software engineering. It targets roughly 180 tokens per second, with short-context responses reaching up to about 260 tokens per second for rapid coding iterations.
Kimi K2.7 Code
moonshotai/kimi-k2.7-codeKimi K2.7 Code is Moonshot AI's coding-focused agentic model built for long-horizon software engineering workflows. It supports native image input, tool calling, and forced thinking mode; instant/non-thinking mode is not supported.
Kimi Latest
moonshotai/kimi-latestCompatibility alias that routes to the newest Kimi model. Currently routes to Kimi K3.
Kimi K2.6
moonshotai/kimi-k2.6Kimi K2.6 is an open-source, native multimodal agentic model built for long-horizon coding, coding-driven design, and large-scale task orchestration. It can turn simple prompts and visual inputs into production-ready interfaces and full-stack workflows, and is designed to coordinate complex multi-agent plans with thousands of steps across code, documents, and spreadsheets.
Kimi K2.5
moonshotai/kimi-k2.5Kimi K2.5 is Moonshot AI's native multimodal model built on Kimi K2 with ~15T mixed visual and text tokens, delivering strong general reasoning, visual coding, and agentic tool-calling. This route uses instant (non-thinking) mode for faster responses.