DeepSeek-V4.1-Flash

Definition

552B-parameter multimodal mixture-of-experts model from deepseek targeting long-horizon agent workloads with aggressive kv-cache compression. API ID: deepseek-flash.

Key Points

• 2026-10-17: DeepSeek-V4.1-Flash-1017 checkpoint — API deepseek-flash routes to 1017; Artificial Analysis Intelligence Index 45 (#2); vendor benchmark uplifts on Terminal-Bench, ProgramBench, HLE (2026-10-19-deepseek-v41-flash-1017-max, 2026-10-19-deepseek-v41-flash-1017-huggingface)

• 2026-09: CED architecture — 8B active prefill, 16B decode; 1M token context (2026-09-17-deepseek-v41-flash-huggingface)

• compressed-sparse-attention CSA2 + FP4 KV: 890 bytes/token global cache (~1/4 V4-Flash HBM, ~1/8 SSD via SWA Bounded Replay)

• Agent benchmarks (harness-dependent): Terminal-Bench 2.1 90.6, DeepSWE 74.2, CyberGym 88.1

• V4-Pro requests route to V4.1-Flash from Sept 14, 2026 (2026-09-17-deepseek-v41-flash-api-docs)

Related

• deepseek
• kv-cache
• mixture-of-experts
• causal-encoder-decoder
• compressed-sparse-attention
• llm-inference
• coding-agents

Sources

• 2026-09-17-deepseek-v41-flash-kv-cache-compression
• 2026-09-17-deepseek-v41-flash-huggingface
• 2026-09-17-deepseek-v41-flash-api-docs
• 2026-09-17-deepseek-v41-flash-deepseekagent