UK | EN |
LIVE
Технології 🇺🇸 США

DeepSeek V4 Pro tops global bang-for-buck ranking after 75% price cut

South China Morning Post Vincent Chow 0 переглядів 2 хв читання
DeepSeek V4 Pro tops global bang-for-buck ranking after 75% price cut
AdvertisementArtificial intelligenceTechTech TrendsDeepSeek V4 Pro tops global bang-for-buck ranking after 75% price cut

Model ranks among world’s best for intelligence per dollar thanks to now permanent price cut, according to Artificial Analysis

2-MIN READ2-MIN Listen
On Saturday, DeepSeek made a 75 per cent price cut for the V4 Pro model permanent. Photo: AFP
Vincent ChowPublished: 8:00pm, 24 May 2026Chinese artificial intelligence start-up DeepSeek’s latest flagship model has been ranked as one of the world’s best on an intelligence-per-dollar basis, far exceeding those from US heavyweights OpenAI and Anthropic in cost-efficiency after a 75 per cent promotional price cut was made permanent.The Hangzhou-based start-up announced the permanent price cut for its V4 Pro model on Saturday, a month after it released the long-awaited V4 generation, which comprises the flagship V4 Pro model and its lighter variant, V4 Flash.

The move reflects how Chinese companies are competing in the global AI race through a different approach from their US counterparts, which offer cutting-edge model capabilities at premium prices.

Advertisement

According to third-party benchmark firm Artificial Analysis, the price cut places V4 Pro at the global frontier for cost-efficiency, based on an assessment of the amount of “intelligence” one can extract per dollar.

Amid a global compute supply crunch that has pushed up prices for the most advanced models, the bang-for-the-buck method of comparing different AI models – as opposed to ranking their raw intelligence – has gained popularity in recent months.

Advertisement

DeepSeek’s official application programming interface (API) price for V4 Pro is now as low as US$0.0036 per 1 million cached input tokens, and US$0.87 per 1 million output tokens.

This means that the model requires US$268 to run Artificial Analysis’ Intelligence Index benchmark tests, a composite benchmark combining several performance metrics. OpenAI and Anthropic’s latest flagship models, GPT-5.5 and Claude Opus 4.7, cost about 12 and 19 times more, respectively, to perform the same task.AdvertisementSelect VoiceSelect Speed0.8x0.9x1.0x1.1x1.2x1.5x1.75x00:0000:001.00x
Поділитися

Схожі новини