DeepSeek v4.1 Flash – Artificial Analysis (artificialanalysis.ai)

DeepSeek has announced the release of its latest model, DeepSeek V4.1 Flash (Reasoning, Max Effort), which hit the market in September 2026. This open weights model is notable for its substantial intelligence, scoring 40 on the Artificial Analysis Intelligence Index, significantly surpassing the median score of 18. Alongside its high performance, the model integrates multimodal capabilities, supporting both text and image inputs with a remarkable processing context window of 1 million tokens, equivalent to approximately 1500 A4 pages of content. The significance of DeepSeek V4.1 Flash lies in its competitive pricing and exceptional speed; it processes output at 190 tokens per second and offers input token costs at $0.30 per million, making it an appealing choice for developers and businesses alike. The model effectively balances cost and performance, generating 250 million tokens during evaluations while remaining above average in verbosity compared to similar models. As AI applications become increasingly complex and diverse, this model's ability to tackle reasoning tasks efficiently positions it as a valuable tool for various sectors, enhancing capabilities in analysis and interaction across different data formats.