DeepSeek has made its 75% V4-Pro API price cut permanent, sharply undercutting OpenAI and Anthropic and escalating pressure on premium closed-model AI economics.
DeepSeek has intensified the open-source AI pricing war by permanently locking in a 75% cut on its V4-Pro API pricing, turning what was initially a temporary promotion into a long-term market shift that directly pressures premium AI providers OpenAI and Anthropic.
The revised pricing, previously scheduled to expire on May 31, now places V4-Pro between $0.0035 and $0.83 per million tokens, down from $0.0145 to $3.48. In comparison, OpenAI’s GPT-5.5 costs $5 per million input tokens and $30 per million output tokens, while Anthropic’s Claude Opus 4.7 charges $5 input and $25 output.
The move significantly alters developer economics. For workloads involving 100 million output tokens monthly, Claude Opus 4.7 costs roughly $2,500 and GPT-5.5 around $3,000, while DeepSeek V4-Pro comes in near $348 — making it nearly seven times cheaper than Anthropic and almost nine times cheaper than OpenAI.
The development signals growing momentum for lower-cost open AI alternatives that are increasingly viable for production workloads. As article author Tobi Opeyemi Amure noted, “Promotional discounts come with an expiration date that competitors can simply outwait. Permanent discounts do not.”
Reuters also reported that DeepSeek did not disclose whether the lower pricing was enabled by expanded availability of Huawei Ascend 950 chips, highlighting the growing role of non-Nvidia AI infrastructure in reducing inference costs.
