OpenAI released GPT-5.6 Sol, a new model that delivers near-parity performance with Anthropic's Claude Fable 5 at a fraction of the cost. Sol scores 59 points on the Artificial Analysis Intelligence Index, trailing Fable 5 by just one point. The pricing gap is substantial: Sol costs $1.04 per task compared to Fable 5's $3.12, representing a 67 percent price reduction.
The competitive pressure extends beyond raw benchmark performance. Sol demonstrates superior capabilities in agentic coding tasks, outperforming every competitor in this specialized domain. This strength matters because autonomous coding agents represent a high-value application area where enterprises increasingly direct AI spending.
The release reflects an ongoing pattern in large language model competition: performance convergence paired with aggressive pricing. As models cluster around similar benchmark scores, cost becomes the primary differentiator. OpenAI's pricing strategy positions Sol as the default choice for cost-conscious customers who need near-frontier performance without paying premium rates.
For Anthropic, the timing creates strategic headwinds. Fable 5 maintains a marginal performance lead, but Sol's one-point gap combined with one-third pricing dilutes that advantage significantly. Enterprise procurement teams evaluating models face a straightforward calculation: spend three times more for one additional benchmark point, or accept Sol's trade-off and redirect savings elsewhere.
The agentic coding advantage adds another layer. As autonomous agents become central to enterprise AI deployments, Sol's coding capabilities give it practical advantages that benchmarks alone don't capture. This specialization could matter more than aggregate scores in real-world adoption.
This development accelerates the commoditization trend in the frontier model market. OpenAI and Anthropic now compete primarily on cost and specific task performance rather than across-the-board capability gaps. Smaller