OpenAI's GPT-6.1 Sol: Near-Astra Power at a Fraction of the Cost

OpenAI's GPT-6.1 Sol: Near-Astra Power at a Fraction of the Cost

OpenAI has officially launched GPT-6.1 Sol, an upgraded iteration of its mid-tier model that promises to deliver near-flagship intelligence without the flagship price tag. Announced at DevDay 2026 in San Francisco, the model is positioned as a 'daily workhorse' for developers and enterprises, offering a compelling balance of capability and cost efficiency.

The new model is available starting today across ChatGPT Work, Codex, and the OpenAI API, with plans to expand to Ultrafast variants in the coming days. This launch comes at a strategic time, as OpenAI looks to solidify its position in the increasingly competitive AI market.

The Price-Performance Equation

GPT-6.1 Sol is priced at $2 per million input tokens and $10 per million output tokens, a significant reduction from GPT-6 Astra's $10/$50 pricing. More striking is the cached input cost: just $0.10 per million tokens, which is 95% lower than standard input pricing and 50% lower than GPT-6 Sol's cached input rate.

This aggressive pricing strategy is designed to encourage developers to build more capable agents that can reuse context across requests without incurring prohibitive costs. For long-horizon tasks, this could be a game-changer in terms of operational economics.

Benchmark Performance: Closing the Gap

Independent testers at Artificial Analysis largely confirm OpenAI's claims, showing Sol trailing Astra by just one point on overall intelligence metrics while costing less than a quarter as much to run. The model shows particular strength in several key areas:

• Coding: On DeepSWE v1.1, GPT-6.1 Sol matches GPT-6 Astra at roughly one-fifth the cost, and outperforms GPT-6 Sol's best score by 6.4 percentage points at lower reasoning effort.
• Professional Work: On GDP.pdf, it scores higher than Opus 5.5 with fallbacks at less than half the cost per task, approaching Astra's state-of-the-art performance.
• Computer Use: On OSWorld 2.0, it outperforms GPT-6 Sol by seven points at max reasoning effort, coming within 2.1 points of Astra at one-seventh the cost per task.
• Scientific Research: On Terminal-Bench Science 0.1, it more than doubles GPT-6 Sol's score at max effort, costing $5.47 per task versus $23.80 for Astra.

Factuality and Safety Improvements

GPT-6.1 Sol also demonstrates improved factual accuracy. At low reasoning effort, the share of responses containing factual errors dropped from 11.4% to 7.7%, a 32% reduction. Across tested settings, its error rate remains within 1.9 percentage points of Astra's, at less than one-fifth the cost per task.

In safety evaluations, GPT-6.1 Sol shows lower failure rates than GPT-6 Sol on transparency about broken tools and respecting explicit restrictions. It also matches Astra and Sol in avoiding attempts to bypass automated safety reviewers, according to the system card addendum.

Strategic Positioning and Market Impact

The launch of GPT-6.1 Sol comes amid a flurry of announcements at DevDay, including new 'Dots' AI agents, a $500/month ChatGPT Pro tier, and the upcoming GPT-6 Astra Ultrafast. OpenAI CEO Sam Altman described Sol as 'smarter than Astra in some ways,' emphasizing its role as a cost-effective alternative for everyday tasks.

This move is widely seen as a response to competitive pressure from models like Claude Opus 5.5, which launched with lower costs. By offering near-Astra performance at a fraction of the price, OpenAI is making a clear play for developer adoption and enterprise scale.

Availability and Future Outlook

GPT-6.1 Sol is now available to all Plus, Pro, Business, Enterprise, and Edu users in ChatGPT Work and Codex. Developers can access it via the API as gpt-6.1-sol. In the coming days, Ultrafast variants with up to 8x faster token generation will be offered in Codex.

As AI models continue to commoditize, the focus shifts to efficiency and cost-effectiveness. GPT-6.1 Sol represents a significant step in that direction, offering a compelling option for teams that need high intelligence without the premium price.