GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price

The cost to use AI for coding tasks falls by up to 81%, making advanced AI code generation more accessible.

OpenAI has released GPT-6.1 Sol, a new model that offers comparable performance to its predecessors on real-world engineering tasks for a significantly lower cost. The FrontierCode 1.1 benchmark, which assesses code quality and mergeability, shows GPT-6.1 Sol achieving a score of 60.4%. This is on par with GPT-6 Sol's 60.7% and GPT-5.6 Sol's 60.6%. Crucially, GPT-6.1 Sol operates at a much reduced price point; at medium reasoning effort, it costs $0.31 per task, an 81% reduction compared to GPT-6 Sol's $1.66. Across various effort levels, GPT-6.1 Sol is 44% to 57% cheaper than GPT-6 Sol while maintaining performance within one percentage point. Notably, at low effort, GPT-6.1 Sol achieves a 58.1% score for $0.21, outperforming GPT-6 Sol's 50.5% at the same setting, making it the top-scoring model under $0.30 per task on the FrontierCode leaderboard.

The introduction of GPT-6.1 Sol signifies a strategic move towards democratizing access to advanced AI coding capabilities by drastically cutting operational costs. While its benchmark performance is nearly identical to higher-priced models, its affordability at lower reasoning efforts is particularly impactful, positioning it as the leading choice for budget-conscious developers and businesses. This pricing strategy could accelerate the adoption of AI coding assistants in a wider range of projects, from small-scale experiments to enterprise-level development, by removing a significant financial barrier.

This development suggests a market trend where incremental performance gains are being traded for substantial cost efficiencies, prioritizing accessibility and broader market penetration over bleeding-edge but expensive capabilities. Companies utilizing AI for code generation will now have a more compelling economic rationale to integrate these tools more deeply into their workflows. Future developments will likely focus on further cost reductions and performance optimizations at lower effort settings, potentially widening the gap between premium and mainstream AI coding solutions.