← Back to Model Beat
Models·3d ago·all news from October 16, 2026

Deepseek V4.1 Flash 1017 beats larger rivals in agentic tasks while once again massively cutting per-task costs

DeepSeek has released the V4.1-Flash-1017 model, which the company reports achieves higher benchmark scores and faster task completion times compared to its predecessor. Beyond the model update, DeepSeek is providing access to its internal search engine via new APIs priced at $0.75 per 1,000 calls. These developments represent a strategic shift toward lower-cost, high-efficiency AI agents, allowing developers to integrate cheaper automated search and reasoning capabilities into their applications.

ModelsDeepSeek V4.1 Flash 1017DeepSeek V4.1 Flash

Covered by 12 sources · 14 articles

Related stories

ModelsNew Deepseek model V4.1-Flash cuts memory needs for AI agentsSep 8 · 30 sourcesModelsDeepSeek 4.1 Flash Uncensored, Abliterated Builds Hit HuggingFace Hours After ReleaseSep 13 · 8 sourcesModelsLWiAI Podcast #258 - Opus 5.5, Sol and Luna, Muse, DeepSeek-V4.1-Flash, XiOct 3ModelsDeepSeek-V4.1-Flash: Pushing the Limits of KV Cache CompressionSep 18