DeepSeek V4 Launches: 1.6T-Parameter Open-Weight Model Hits 80.6% SWE-bench at 1/6th the Cost of Closed Rivals

simonwillison.net·high signal

DeepSeek released V4-Pro (1.6T total params, 49B active, MIT license) and V4-Flash (284B total, 13B active) on April 24 with native 1M-token context. V4-Pro scores 80.6% on SWE-bench Verified — within 0.2 points of Claude Opus 4.6 — and leads Claude on Terminal-Bench 2.0 (67.9% vs 65.4%) and LiveCodeBench (93.5% vs 88.8%). The key architectural innovation is Hybrid Attention combining Compressed Sparse and Heavily Compressed attention, requiring only 27% of inference FLOPs and 10% of KV cache vs. V3.2 at 1M context. Pricing: V4-Flash at $0.14/$0.28 per million tokens is the cheapest model in its tier.