DeepSeek V4 Pro Launches with Significant Cost Advantage
- DeepSeek’s V4-Pro model features a massive 1.6 trillion parameters, making it the largest open-source model available.
- The cost for using V4-Pro is $1.74 per million input tokens and $3.48 per million output tokens, significantly cheaper than its competitors.
- DeepSeek trained the V4 on Huawei Ascend chips, circumventing U.S. export restrictions, and plans to further reduce costs with new supernodes in late 2026.
- V4-Flash offers a practical alternative with lower parameters but comparable reasoning performance at even lower costs.
DeepSeek has released its V4-Pro and V4-Flash models, offering substantial cost savings compared to Western counterparts like GPT-5.5 Pro and Claude Opus models. The strategic use of Huawei Ascend chips allows DeepSeek to bypass U.S.-imposed chip bans, maintaining competitive pricing in the AI market.
The introduction of these models could shift enterprise AI strategies due to their affordability and efficiency, particularly for large-scale operations like document processing or code generation pipelines. Source
