Kimi K2 Base
Zero-eval
#1C-Eval
#1MMLU-redux-2.0
#1TriviaQA
+2 more
by Moonshot AI
+
+
+
+
About
Kimi K2 Base was created as the foundation model in the K2 series, designed to serve as a starting point for fine-tuning and customization. Built to provide strong base capabilities for domain-specific applications, it enables developers to build specialized solutions on Moonshot's architecture.
+
+
+
+
Timeline
AnnouncedJul 11, 2025
ReleasedJul 11, 2025
+
+
+
+
Specifications
Training Tokens15.5T
+
+
+
+
License & Family
License
MIT
Performance Overview
Performance metrics and category breakdown
Overall Performance
13 benchmarks
Average Score
69.2%
Best Score
92.5%
High Performers (80%+)
6+
+
+
+
All Benchmark Results for Kimi K2 Base
Complete list of benchmark scores with detailed information
| C-Eval | text | 0.93 | 92.5% | Self-reported | |
| GSM8k | text | 0.92 | 92.1% | Self-reported | |
| MMLU-redux-2.0 | text | 0.90 | 90.2% | Self-reported | |
| MMLU | text | 0.88 | 87.8% | Self-reported | |
| TriviaQA | text | 0.85 | 85.1% | Self-reported | |
| EvalPlus | text | 0.80 | 80.3% | Self-reported | |
| CSimpleQA | text | 0.78 | 77.6% | Self-reported | |
| MATH | text | 0.70 | 70.2% | Self-reported | |
| MMLU-Pro | text | 0.69 | 69.2% | Self-reported | |
| GPQA | text | 0.48 | 48.1% | Self-reported |
Showing 1 to 10 of 13 benchmarks
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+