Quality
#31/6717.3
Overall quality score
🇨🇳Moonshot AIOpen sourceMoonshotAI: Kimi K2 Thinking.
See how MoonshotAI: Kimi K2 Thinking compares with other broadly useful AI models across quality, speed, price, context, and practical capabilities.
Speed
#31/6779 t/s
Output tokens per second
Input price
#40/67$0.60/1M
Per 1M tokens
Output price
#41/67$2.50/1M
Per 1M tokens
Context
#37/67262.1K
Content read at once
Comparison summary
MoonshotAI: Kimi K2 Thinking ranks #31 of 67 for overall quality and #31 for response speed in the current catalogue.
Its listed input price is $0.60/1M, output price is $2.50/1M, and it can work with up to 262.1K of context at once.
Its strongest areas in the current data are coding, everyday help, math and logic.
Practical specifications
- Company
- Moonshot AI
- Reasoning
- Yes
- Open weights
- Yes
- Input
- Text
- Output
- Text
- Context window
- 262.1K
MoonshotAI: Kimi K2 Thinking technical benchmarks.
Named evaluation records for technical comparison. Results retain their original benchmark names and sources.
| Benchmark | Area | Kimi K2 Thinking | Source |
|---|---|---|---|
| Intelligence Index | overallQuality | 17.3 | OpenRouter / artificial-analysis |
| Coding Index | coding | 21 | OpenRouter / artificial-analysis |
| Agentic Index | agentic | 1.8 | OpenRouter / artificial-analysis |
| GPQA | GPQA Diamond | 71.3 | OpenRouter model benchmarks |
| Humanity's Last Exam | HLE | 9.5 | OpenRouter model benchmarks |
| IFBench | IFBench | 62.8 | OpenRouter model benchmarks |
| τ²-Bench Telecom | τ²-Bench Telecom | 25.4 | OpenRouter model benchmarks |
| AA-LCR | AA-LCR | 52.7 | OpenRouter model benchmarks |
| GDPval-AA | GDPval-AA | 0 | OpenRouter model benchmarks |
| CritPt | CritPt | 0 | OpenRouter model benchmarks |
| SciCode | SciCode | 33 | OpenRouter model benchmarks |
| Terminal-Bench Hard | Terminal-Bench Hard | 6.8 | OpenRouter model benchmarks |
| AA-Omniscience Accuracy | AA-Omniscience Accuracy | 15.7 | OpenRouter model benchmarks |
| AA-Omniscience Non-Hallucination Rate | AA-Omniscience Non-Hallucination Rate | 41.1 | OpenRouter model benchmarks |
| MMLU-Pro | Knowledge and reasoning | 84.6 | OpenEvals/leaderboard-data |
| SWE-bench Verified | Coding | 71.3 | OpenEvals/leaderboard-data |
| Terminal-Bench | Agentic coding | 35.7 | OpenEvals/leaderboard-data |
Quality benchmarks
See where MoonshotAI: Kimi K2 Thinking stands.
The darker bar marks MoonshotAI: Kimi K2 Thinking; the lighter bars provide context from other leading models.
Highlights
Quality
Overall capability · Higher is better
Speed
Output tokens per second · Higher is better
Input price
USD per 1M tokens · Lower is better