Moonshot AI model product
Kimi K2
Kimi K2 0905 is the September update of Kimi K2 0711.
Updated Aug 17, 2026. Default version: Kimi K2-Instruct-0905
LLMBoard Score36.6Kimi K2-Instruct-0905
Coverage100%26 benchmark families
Context windowN/ATokens
Official input price$0.60Moonshot AI API
Kimi K2 Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Kimi K2 Benchmark Results
Benchmark scores for Kimi K2-Instruct-0905.
Sort by
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkACEBench | Score76.5% | Rank02 | Participants2 | Percentile0.0% | EvidenceC | Evaluated |
| BenchmarkAutoLogi | Score89.5% | Rank02 | Participants2 | Percentile0.0% | EvidenceC | Evaluated |
| BenchmarkCNMO 2024 | Score74.3% | Rank02 | Participants3 | Percentile50.0% | EvidenceC | Evaluated |
| BenchmarkPolyMath-en | Score65.1% | Rank02 | Participants2 | Percentile0.0% | EvidenceC | Evaluated |
| BenchmarkMultiPL-E | Score85.7% | Rank05 | Participants13 | Percentile66.7% | EvidenceC | Evaluated |
| BenchmarkZebraLogic | Score89.0% | Rank06 | Participants8 | Percentile28.6% | EvidenceC | Evaluated |
| BenchmarkMATH-500 | Score97.4% | Rank07 | Participants32 | Percentile80.7% | EvidenceC | Evaluated |
| BenchmarkOJBench | Score27.1% | Rank09 | Participants9 | Percentile0.0% | EvidenceC | Evaluated |
| BenchmarkLiveBench | Score76.4% | Rank10 | Participants38 | Percentile75.7% | EvidenceC | Evaluated |
| BenchmarkAider-Polyglot | Score60.0% | Rank13 | Participants22 | Percentile42.9% | EvidenceC | Evaluated |
| BenchmarkMMLU | Score89.5% | Rank14 | Participants101 | Percentile87.0% | EvidenceC | Evaluated |
| BenchmarkMulti-Challenge | Score54.1% | Rank15 | Participants29 | Percentile50.0% | EvidenceC | Evaluated |
| BenchmarkIFEval | Score89.8% | Rank17 | Participants67 | Percentile75.8% | EvidenceC | Evaluated |
| BenchmarkMMLU-Redux | Score92.7% | Rank17 | Participants48 | Percentile66.0% | EvidenceC | Evaluated |
| BenchmarkTau2 Airline | Score56.5% | Rank17 | Participants23 | Percentile27.3% | EvidenceC | Evaluated |
| BenchmarkTau2 Retail | Score70.6% | Rank21 | Participants26 | Percentile20.0% | EvidenceC | Evaluated |
| BenchmarkSuperGPQA | Score57.2% | Rank22 | Participants34 | Percentile36.4% | EvidenceC | Evaluated |
| BenchmarkTerminal-Bench | Score25.0% | Rank23 | Participants25 | Percentile8.3% | EvidenceC | Evaluated |
| BenchmarkSimpleQA | Score31.0% | Rank25 | Participants47 | Percentile47.8% | EvidenceC | Evaluated |
| BenchmarkTau2 Telecom | Score65.8% | Rank28 | Participants35 | Percentile20.6% | EvidenceC | Evaluated |
| BenchmarkHMMT 2025 | Score38.8% | Rank30 | Participants33 | Percentile9.4% | EvidenceC | Evaluated |
| BenchmarkSWE-bench Multilingual | Score47.3% | Rank34 | Participants38 | Percentile10.8% | EvidenceC | Evaluated |
| BenchmarkAIME 2024 | Score69.6% | Rank42 | Participants53 | Percentile21.1% | EvidenceC | Evaluated |
| BenchmarkLiveCodeBench | Score53.7% | Rank42 | Participants75 | Percentile44.6% | EvidenceC | Evaluated |
| BenchmarkMMLU-Pro | Score81.1% | Rank50 | Participants134 | Percentile63.2% | EvidenceC | Evaluated |
| BenchmarkSWE-Bench Verified | Score65.8% | Rank78 | Participants111 | Percentile30.0% | EvidenceC | Evaluated |
| BenchmarkHumanity's Last Exam | Score4.7% | Rank98 | Participants99 | Percentile1.0% | EvidenceC | Evaluated |
| BenchmarkAIME 2025 | Score49.5% | Rank104 | Participants115 | Percentile9.7% | EvidenceC | Evaluated |
| BenchmarkGPQA | Score75.1% | Rank111 | Participants239 | Percentile53.8% | EvidenceC | Evaluated |
Kimi K2 Arena Results
Preference and agent-evaluation results for the default version.
No Arena results
The default version does not have a matching Arena result yet.
Kimi K2 Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $0.60 input, $2.5 output per 1M
- Official provider
- Moonshot AI
- Lowest third-party
- From $1 input, $3 output per 1M via Hugging Face
- Tracked offerings
- 2
Sort by
Provider | Provider model ID | Region | Input / 1M | Output / 1M | Context | Updated |
|---|
| ProviderHugging Face | Provider model IDmoonshotai/Kimi-K2-Instruct-0905 | Regionglobal | Input / 1M$1 | Output / 1M$3 | Context262.1K | Updated |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Kimi K2 Runtime Performance
Provider-specific output speed and catalog latency for Kimi K2-Instruct-0905. Runtime does not affect the capability score.
No runtime data
Browse runtime rankingsNo provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Kimi K2 Specifications
Technical details for the model's default version.
- Version
- Kimi K2-Instruct-0905
- Released
- Sep 5, 2025
- Knowledge cutoff
- Unknown
- Parameters
- 1T
- Context window
- N/A
- Max output
- N/A
- Inputs
- text
- Outputs
- text
- Open weights
- No
- License
- MIT
Kimi K2 Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
Sort by
Version | Released | LLMBoard | Parameters | Context | Max output | Open weights | License |
|---|
| VersionKimi K2 0905 | Released | LLMBoardN/A | Parameters1T | Context262.1K | Max output262.1K | Open weightsNo | LicenseProprietary |
| VersionKimi K2-Instruct-0905 | Released | LLMBoard36.6 | Parameters1T | ContextN/A | Max outputN/A | Open weightsNo | LicenseMIT |
| VersionKimi K2 Base | Released | LLMBoardN/A | Parameters1T | ContextN/A | Max outputN/A | Open weightsNo | LicenseMIT |
| VersionKimi K2 Instruct | Released | LLMBoardN/A | Parameters1T | Context200K | Max output200K | Open weightsNo | LicenseMIT |
Kimi K2 vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to Kimi K2
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
MA
Kimi k1.5
Moonshot AI
MA
Kimi K2 Thinking
Moonshot AI
MA
Kimi K2.5
Moonshot AI
MA
Kimi K2.7 Code
Moonshot AI
MA
Kimi K2.6
Moonshot AI
MA
Kimi K3
Moonshot AI
What is Kimi K2?
Key information about Kimi K2 and its available data.
Kimi K2 0905 is the September update of Kimi K2 0711. It is a large-scale Mixture-of-Experts (MoE) language model developed by Moonshot AI, featuring 1 trillion total parameters with 32 billion active per forward pass. It supports long-context inference up to 256k tokens, extended from the previous 128k.
This update improves agentic coding with higher accuracy and better generalization across scaffolds, and enhances frontend coding with more aesthetic and functional outputs for web, 3D, and related tasks. The model is trained with a novel stack incorporating the MuonClip optimizer for stable large-scale MoE training.
Data as of 2026-08-17.
FAQ
Common questions about Kimi K2.
When was Kimi K2 released?
Kimi K2's default version was released on Sep 5, 2025.
How much does Kimi K2 cost?
Kimi K2's official API price is $0.60 per million input tokens and $2.5 per million output tokens via Moonshot AI. The lowest tracked third-party offer starts at $1 input and $3 output via Hugging Face.
Who created Kimi K2?
Kimi K2 was created by Moonshot AI.
What is the context window for Kimi K2?
A context window is not available for the default version.
Is Kimi K2 open weight?
No. The default version is not marked as having publicly available weights.
How many API providers offer Kimi K2?
2 provider offerings are linked to the default version.
What models should I compare Kimi K2 with?
Nearby ranked alternatives include Claude Sonnet 3.7, Gemma 4 12B, Nemotron 3.5 Lightning.