Artificial Analysis: GPT-6.1 Sol trails GPT-6 Astra by 1 point at under a quarter of the cost

Artificial Analysis says GPT-6.1 Sol replaces GPT-6 Sol after seven days. On its Intelligence Index, the model gains 4 points versus GPT-6 Sol and 5 versus GPT-5.6 Sol, finishing 1 point below GPT-6 Astra.

The reported gains are 4 points on AA-Briefcase v1.1, 5 on GDPval-AA v2.1, 12 on Terminal-Bench 4.0, 5 on Humanity’s Last Exam, and 6 on GDP.pdf. AA-Omniscience Accuracy rises 8 points, while the hallucination rate falls from 60% to 54%. On the Coding Agent Index at max effort, GPT-6.1 Sol gains 3 points over GPT-6 Sol and sits 2 points below GPT-6 Astra.

List pricing matches GPT-6 Sol at $2 per million input tokens and $10 per million output tokens. The cache-read discount rises from 90% to 95%, which Artificial Analysis says makes the blended price for agentic workloads slightly lower than GPT-6 Sol and adds to GPT-6 Sol’s earlier 50% discount versus GPT-5.6 Sol. At max effort, cost per Intelligence Index task is $0.72, less than a quarter of GPT-6 Astra’s $3.26, 31% below GPT-6 Sol’s $1.05, and 64% below GPT-5.6 Sol’s $1.99. Artificial Analysis says every effort level extends the cost-efficiency Pareto frontier. GPT-6.1 Sol uses about 10–30% more output tokens than GPT-6 Sol, but its low and medium effort levels are Pareto optimal for token efficiency.