DeepSeek V4.1 Flash combines strong benchmarks with competitive pricing

DeepSeek V4.1 Flash 1017 (Max) scores 32 on the Artificial Analysis Intelligence Index, above the median of 23 among open-weights models of similar size. Released on October 17, 2026, the reasoning model supports text and image input and generates text output. Its benchmark results place it among the leading models in its comparison group.

Pricing through DeepSeek’s API is $0.30 per 1M input tokens and $1.20 per 1M output tokens, below the respective comparison medians of $0.50 and $2.10. The average cost per Intelligence Index task is $0.60. Output speed reaches 175.6 tokens per second, with a time to first token of 1.19s. However, the model generated 170M output tokens during Intelligence Index evaluation, compared with a median of 100M, indicating substantially greater verbosity.

The model has a context window of 1.0M tokens and uses a mixture-of-experts architecture with 552 billion total parameters and 16 billion active parameters during inference. Its weights are publicly available for self-hosting under the MIT license, which permits commercial use. API access is available through 24 providers, with pricing potentially varying by provider.