Home / Technology / DeepSeek AI V4.1-Flash: Faster, Sharper, Open Source
DeepSeek AI V4.1-Flash: Faster, Sharper, Open Source
19 Oct
Summary
- New AI model V4.1-Flash finishes tasks faster and writes concise responses.
- Benchmarks show significant improvements across multiple testing platforms.
- Search & Fetch APIs are now open to all users at $0.75 per 1,000 calls.
DeepSeek has released an updated version of its V4.1-Flash AI model, named V4.1-Flash-1017, which offers accelerated task completion and more concise responses. This new iteration shows substantial performance gains, achieving scores of 42.6 on Humanity's Last Exam (up from 36.8) and 42.3 on Terminal-Bench 4.0 (up from 31.2).
The company has also made its Search & Fetch APIs available to all users, with each endpoint costing $0.75 per 1,000 calls. This release enhances DeepSeek's competitive standing against industry rivals. The model is accessible via DeepSeek's website, app, and API, and has also been released as open source on Hugging Face and ModelScope.
DeepSeek is currently in the process of raising substantial funding, potentially securing over $12 billion. This investment round values the company at a minimum of 500 billion yuan, with significant backing from major investors. Meanwhile, the company experienced a server outage lasting approximately five hours last week.
