+
+
+
+
+
+
+
+
VS
VS
VS
VS
VS
VS
DeepSeek-V2.5 vs Kimi K2 0905
Comprehensive side-by-side LLM comparison
Kimi K2 0905 leads with 9.9% higher average benchmark score. Kimi K2 0905 offers 507.9K more tokens in context window than DeepSeek-V2.5. DeepSeek-V2.5 is $2.68 cheaper per million tokens. Overall, Kimi K2 0905 is the stronger choice for coding tasks.
+
+
+
+
DeepSeek
DeepSeek-V2.5 was developed as an enhanced iteration of the DeepSeek-V2 architecture, designed to incorporate improvements in model quality and efficiency. Built to advance the DeepSeek foundation model series, it provides refined capabilities for general-purpose language understanding and generation tasks.
+
+
+
+
Moonshot AI
Kimi K2 was introduced as the second generation of Moonshot's language model family, designed to provide enhanced capabilities across language understanding and generation. Built with architectural improvements and expanded training, it represents a significant advancement in Moonshot's model offerings.
1 year newer
DeepSeek-V2.5
DeepSeek
2024-05-08
Kimi K2 0905
Moonshot AI
2025-09-05
+
+
+
+
+
+
+
+
+
+
+
+
Pricing Comparison
Cost per million tokens (USD)
+
+
+
+
DeepSeek-V2.5
Input:$0.14
Output:$0.28($2.68 cheaper)
+
+
+
+
Kimi K2 0905
Input:$0.60
Output:$2.50
Performance Metrics
Context window and performance specifications
Average performance across 3 common benchmarks
+
+
+
+
DeepSeek-V2.5
Average Score:81.4%
+
+
+
+
Kimi K2 0905
Average Score:91.3%(+9.9%)
Provider Availability & Performance
Available providers and their performance metrics
+
+
+
+
DeepSeek-V2.5
3 providers
DeepInfra
Throughput: 63 tok/s
Latency: 0.5ms
DeepSeek
Throughput: 100 tok/s
Latency: 0.5ms
Hyperbolic
Throughput: 100 tok/s
Latency: 0.5ms
+
+
+
+
+
+
+
DeepSeek-V2.5
Avg Score:81.4%
Providers:3
+
+
+
+
Kimi K2 0905
Avg Score:91.3%(+9.9%)
Providers:2