MoonshotAI: Kimi K3
by Moonshot AItext+image+video->text20 endpoints
Neura Intelligence Index
88.2/ 100
Rank
#17—
Confidence
highBenchmarks
8(21% coverage)
Overview
Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI. It is suited for complex coding, knowledge work, and long-horizon agentic workflows, and is particularly strong at...
Capabilities
Text generation
Image understanding
Video understanding
Tool use / Function calling
Structured output
Extended reasoning
Prompt caching
Long context
Top-P sampling
Stop sequences
Deterministic seed
Log probabilities
Modalities
Input
TextImageVideo
Output
Text
Technical Specifications
Context Window
1.0M tokens
Max Output
943.7K tokens
Tokenizer
Other
Uptime (24h)
9991.7%
Parameters
2800000000000B
License
kimi_k3Open Weights
Country
🇨🇳CN
Release Date
July 16, 20262mo
Supported Parameters
frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_tokensmin_ppresence_penaltyreasoningreasoning_effortrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p
Benchmark Performance
reasoning
| Benchmark | Score | Source |
|---|---|---|
| GPQA Diamond | 93.5 | verified |
| Humanity's Last Exam | 56 | verified |
multimodal
| Benchmark | Score | Source |
|---|---|---|
| CharXiv Reasoning | 91.3 | verified |
| MMMU-Pro | 81.6 | verified |
agent
| Benchmark | Score | Source |
|---|---|---|
| BrowseComp | 91.2 | verified |
| MCP Atlas | 84.2 | verified |
| Toolathlon | 73.2 | verified |
| APEX Agents | 37.6 | verified |
Performance Over Time
Pricing
Per 1M tokens
Input$2.10
Output$10.95
Cache read$0.23
Added
July 16, 2026
Last synced 9/18/2026