DeepSeek V4 Pro tops cost-efficiency ranking after 75% price cut, beating US rivals
V4 Pro costs 12x less than GPT-5.5 to run identical benchmark tests.
Get AI news that actually matters
One email a day. Zero fluff. Join 10,000+ professionals.
DeepSeek, the Hangzhou-based AI startup, has permanently slashed API prices for its flagship V4 Pro model by 75%, a month after releasing the V4 generation. The price cut places the model at the top of Artificial Analysis' global bang-for-buck ranking, which measures intelligence per dollar. Specifically, V4 Pro now charges $0.0036 per 1 million cached input tokens and $0.87 per 1 million output tokens. Running Artificial Analysis' Intelligence Index benchmark—a composite of performance metrics—costs just $268 with V4 Pro. In contrast, OpenAI's GPT-5.5 costs 12 times more for the same task, and Anthropic's Claude Opus 4.7 demands 19 times more.
This strategic pricing reflects a broader trend in the AI industry: amid a global compute supply crunch that has driven up costs for premium models, the bang-for-buck metric has gained popularity. While US heavyweights focus on cutting-edge capabilities at premium prices, Chinese companies like DeepSeek are competing on cost efficiency without sacrificing performance. The permanent price cut also signals confidence in V4 Pro's capabilities, as DeepSeek makes a long-term bet on volume over margins. For developers and enterprises, this means access to frontier-level AI at a fraction of the cost—potentially accelerating adoption in cost-sensitive applications.
- DeepSeek made a 75% price cut permanent for V4 Pro, now charging $0.0036 per 1M cached input tokens and $0.87 per 1M output tokens.
- V4 Pro tops Artificial Analysis' intelligence-per-dollar ranking, costing $268 to run their benchmark vs. $3,216 for GPT-5.5 and $5,092 for Claude Opus 4.7.
- The move underscores China's AI strategy of competing on affordability amid a global compute crunch, challenging US dominance on cost rather than raw power.
Why It Matters
DeepSeek's aggressive pricing forces US AI leaders to compete on value, potentially democratizing access to frontier models for startups and enterprises.