DeepSeek benchmarks
DeepSeek benchmarks Articles
Browse 3 articles about DeepSeek benchmarks.
DeepSeek V4.1 Flash Specs: KV Cache Compression Explained
DeepSeek V4.1 Flash's model card breaks down its 552B MoE design, 1M context window, and 890-byte KV cache per token in detail.
DeepSeek V4.1 Flash specsDeepSeek architectureKV cache compression
DeepSeek-V4-Pro-0813: What's New and How It Stacks Up
DeepSeek-V4-Pro-0813 adds DSpark speculative decoding and beats its preview on coding, agentic, and tool-use benchmarks versus GLM 5.2, Kimi K3, and Opus 4.8.
DeepSeek V4 ProDeepSeek-V4-Pro-0813DSpark speculative decoding
DeepSeek V4 Flash: The Cheapest Frontier-Level Open Model Yet
DeepSeek V4 Flash jumps to 54% on agentic coding benchmarks and rivals larger models, but harness choice explains much of the gain.
DeepSeek V4 FlashDeepSeek benchmarksopen source LLM