OpenAI GPT-5.6 wins China user praise despite higher cost than local AI rivals
Synopsis
OpenAI's GPT-5.6 debuts three models — Sol, Terra, Luna — with Terra priced at half of GPT-5.5. Yet even at reduced rates, it costs up to 34x more per output token than DeepSeek V4, exposing the widening price gulf between Western and Chinese AI providers.
Key Takeaways
OpenAI launched GPT-5.6 on Thursday, July 10, 2026 , comprising three models: flagship Sol , balanced Terra , and lightweight Luna .
Terra is priced at US$2.50 per million input tokens and US$15 per million output tokens — roughly half the cost of GPT-5.5 , according to OpenAI .
DeepSeek V4 undercuts GPT-5.6 Sol significantly, priced at up to US$0.44 per million input tokens and US$0.87 per million output tokens .
Zhipu AI 's GLM-5.2 charges approximately US$1.40 per million input tokens and US$4.40 per million output tokens .
Li Yitao , co-founder of Canada -based AI start-up Quotaflow , described the launch as part of a broader shift toward cost-efficient enterprise AI.
Chinese users continue to access GPT-5.6 via VPNs and third-party proxies, as OpenAI remains blocked in China .
OpenAI's newly launched GPT-5.6 is drawing positive responses from users in China who access the service through VPNs and third-party proxies, with many citing improved cost-efficiency — even as the American AI model remains pricier than homegrown competitors. The release, made on Thursday, July 10, 2026, introduces three distinct models targeting different performance tiers and budgets.
Three-tier model lineup: Sol, Terra, Luna
The GPT-5.6 family comprises the flagship Sol, the mid-range Terra, and the lightweight Luna. According to OpenAI, Terra delivers performance on par with its predecessor, GPT-5.5, at roughly half the price, while Luna is engineered to offer robust capabilities at a low inference cost.
Sol is priced at US$5 per million input tokens and US$30 per million output tokens. Terra comes in at US$2.50 and US$15 respectively, while Luna — the entry-level option — costs US$1 per million input tokens and US$6 per million output tokens.
Why it matters: enterprise AI shifts toward cost efficiency
The launch reflects a broader industry pivot toward cost-efficient AI systems built for enterprise deployment, according to Li Yitao, a Chinese entrepreneur and co-founder of Canada-based AI start-up Quotaflow. The tiered pricing structure signals that OpenAI is actively competing on value, not just capability.
Despite meaningful price reductions compared with earlier OpenAI products, the flagship and mid-tier GPT-5.6 models still sit well above Chinese alternatives on a per-token basis — a gap that enterprise buyers are increasingly scrutinising.
The competitive backdrop: Chinese rivals undercut on price
Zhipu AI's GLM-5.2 charges approximately US$1.40 per million input tokens and US$4.40 per million output tokens. DeepSeek V4 is priced even lower, at up to US$0.44 per million input tokens and US$0.87 per million output tokens — making it roughly 11 times cheaper on input and more than 34 times cheaper on output than GPT-5.6 Sol.
The stark price differential underscores the intensifying competition in the global large language model market, where Chinese developers have aggressively undercut Western incumbents to capture enterprise market share.
What's next: access barriers and market exposure
Chinese users accessing GPT-5.6 must still navigate the country's internet restrictions, relying on VPNs and proxy services — a structural barrier that limits OpenAI's addressable market in the world's largest internet economy. As domestic rivals continue to narrow the capability gap while maintaining far lower price points, OpenAI's ability to retain its premium positioning among cost-sensitive enterprise buyers in Asia will be closely watched.
Point of View
Whose aggressive per-token economics have forced Western AI leaders to compete on cost — territory where Chinese developers hold a structural advantage. What mainstream coverage often underplays is the access paradox: OpenAI's strongest organic demand signal in China comes from users willing to circumvent national internet controls, yet that same regulatory firewall caps the company's commercial upside in the market. The price gap — DeepSeek V4 at US$0.87 versus GPT-5.6 Sol at US$30 per million output tokens — is too wide to be bridged by performance premiums alone for cost-sensitive enterprise buyers. If this trajectory continues, OpenAI risks being cornered into a high-margin niche even as the broader AI inference market races to the bottom.
NationPress
10 Jul 2026
Frequently Asked Questions
What is OpenAI GPT-5.6 and what models does it include?
GPT-5.6 is OpenAI 's latest large language model release, launched on Thursday, July 10, 2026 . It includes three variants: the flagship Sol , the mid-range Terra , and the lightweight Luna , each targeting different performance and price requirements.
How does GPT-5.6 pricing compare to DeepSeek and Zhipu AI?
GPT-5.6 Sol costs US$30 per million output tokens , compared with US$0.87 for DeepSeek V4 and US$4.40 for Zhipu AI 's GLM-5.2 . Even the entry-level Luna at US$6 per million output tokens remains more expensive than both Chinese rivals.
Why are Chinese users accessing GPT-5.6 despite it being blocked?
Chinese users access GPT-5.6 via VPNs and third-party proxies because OpenAI 's services are blocked in China . Many users reportedly favour the model for its cost-efficiency improvements relative to earlier OpenAI products, even at a higher price than domestic alternatives.
What is GPT-5.6 Terra and how does it compare to GPT-5.5?
Terra is the mid-tier model in the GPT-5.6 family, priced at US$2.50 per million input tokens and US$15 per million output tokens . According to OpenAI , it delivers performance comparable to GPT-5.5 at roughly half the price.
What does GPT-5.6 mean for the US-China AI competition?
The launch highlights the intensifying price war between Western and Chinese AI developers. Chinese rivals like DeepSeek and Zhipu AI have established a significant cost advantage, pressuring OpenAI to reduce prices while still maintaining a premium — a tension that will define enterprise AI adoption trends through 2026 and beyond.
