Simon Willison • 4/24/2026

DeepSeek V4 - almost on the frontier, a fraction of the price

This article covers the release of DeepSeek V4 Pro and V4 Flash, two new Mixture of Experts models with 1M token context. Pro has 1.6T total parameters (49B active), Flash has 284B total (13B active), both under MIT license. The article highlights their competitive performance against frontier models from OpenAI, Google, and Anthropic, while being drastically cheaper—Flash at $0.14/M input and Pro at $1.74/M input. Efficiency improvements reduce FLOPs and KV cache size versus V3.2. Includes benchmark comparisons, pricing table, and SVG generation examples.

0 comments

#Open Weights #Mixture Of Experts #LLM Pricing

$DeepSeek V4 - almost on the frontier, a fraction of the price$