GPT-5.6 Sol nearly matches Fable 5 on aggregated benchmarks at one-third the cost

What happened

OpenAI’s GPT-5.6 Sol has scored 59 points on the Artificial Analysis Intelligence Index, narrowly trailing Anthropic’s Claude Fable 5, which scored 60. Sol achieves this performance at $1.04 per task, roughly one-third the cost of Anthropic’s leading model. Beyond benchmarks, Sol also outperforms competitors in agentic coding tasks, demonstrating stronger capabilities where autonomous code generation and task execution matter most.

Why it matters

GPT-5.6 Sol’s ability to match Anthropic’s top model on a widely recognized aggregated benchmark while dramatically undercutting cost introduces serious pricing pressure in the high-end AI landscape. For businesses and developers, this shifts the economics of deploying advanced AI. Lower-cost access to near state-of-the-art performance can enable more affordable scaling of AI-driven applications, especially in coding and automation workflows where Sol leads.

Anthropic faces direct competition not just on raw capability but also on price-performance balance. This challenges them to either lower costs or differentiate through specialized features. Investors and operators should watch how cost efficiency influences adoption among builders already weighing Claude models.

What to watch next

Tracking how OpenAI scales access to GPT-5.6 Sol across real-world use cases—beyond benchmarks—will reveal if the price advantage holds after volume discounts and integration overhead. Watch Anthropic’s response in pricing, product updates, or new models aimed at reclaiming leadership in agentic tasks and cost efficiency.

Adoption trends among AI builders and enterprise clients will clarify if Sol’s winning formula causes a market shift or triggers wider pricing competition. This squeeze could accelerate commoditization of top-tier AI or drive innovation around specialized agents and workflow automation.

AI Quick Briefs Editorial Desk