DeepSeek V4 Pro and V4 Flash ship under MIT license with 1M-token context — V4 Pro at 1.6T total / 49B active leads Artificial Analysis Index 52 for open weights
DeepSeek released V4 Pro and V4 Flash on April 24, 2026, both MIT-licensed with a 1M-token context. The official model cards on Hugging Face expose both variants: V4-Pro (1.6T total / 49B active) and V4-Flash (284B total / 13B active), both with 1M context and MIT licensing. As of the May 2026 neutral ranking, V4 Pro sits first among open-weight models on the Artificial Analysis Index at 52 and ranks #1 agentic among open weights.
The capability-and-licensing substance is the operational piece. V4 Pro at 1.6T total parameters / 49B active is the largest MoE open-weight model under permissive licensing as of May 2026, with the 1M-token context window matching or exceeding most closed-weight competitors. The MIT license — even more permissive than Apache 2.0 — places zero compliance friction on enterprise adoption. The #1 agentic ranking among open weights means V4 Pro is the most capable open-weight model for tool-using and multi-turn agent workloads, which is the high-value enterprise-deployment segment. V4 Flash at 284B/13B handles the cheaper-inference workload class with the same architectural family.
The competitive context is the open-weight versus closed-weight competitive surface. Mistral Large 3 and Mistral Small 4 shipping under Apache 2.0 is the European-side open-weight commitment at the frontier-flagship tier; DeepSeek V4 Pro under MIT is the Chinese-side equivalent. Claude Opus 4.8's closed-weight release the same week represents the closed-weight frontier. The procurement decision splits on workload risk profile: workloads where capability gap is small and IP sensitivity is high benefit from MIT-licensed open weights; workloads requiring frontier capability across multiple axes still favor closed-weight flagship models. The competitive surface is multi-axis rather than open-versus-closed binary.
Codersera — Best Open-Source LLM May 2026 DeepSeek V4 comparison → · LLM Stats — AI Updates Today May 2026 Latest AI Model Releases → · DeepSeek — V4 Pro V4 Flash MIT release April 24 2026 →