OpenAI has expanded its GPT-6 lineup with two new models—GPT-6 Sol and GPT-6 Luna—delivering significant price reductions while maintaining strong performance across professional work, coding, and computer use tasks.
The new models represent OpenAI’s effort to make frontier AI more accessible. GPT-6 Sol is positioned as the cost-efficient alternative to the flagship GPT-6 Astra, while GPT-6 Luna targets high-volume, lower-complexity workloads.
Pricing That Changes the Game
The pricing overhaul slashes costs by 50% compared to GPT-5.6 promotional rates. GPT-6 Sol now costs $2 per million input tokens and $10 per million output tokens, down from $4 and $20 respectively. GPT-6 Luna drops to $0.10 input and $0.50 output, making it one of the most affordable frontier-tier options available.
“We trained GPT-6 Sol and Luna with similar methods as GPT-6 Astra, bringing the advances behind Astra’s state-of-the-art performance in professional work, factuality, coding, computer use, and alignment to faster, more affordable models,” OpenAI stated in the announcement.
Benchmark Performance
The new models demonstrate impressive efficiency gains. On AutomationBench, which tests business workflows across 47 tools spanning sales, marketing, operations, and finance, GPT-6 Sol at xhigh effort outperforms Claude Opus 5 at max effort at just 9% of the cost.
On coding tasks measured by DeepSWE v1.1, GPT-6 Sol at max effort scores 68.8%—within 1.1 percentage points of Claude Fable 5’s highest score—at approximately 80% lower cost per task. GPT-6 Luna achieves 66.6%, comparable to Claude Opus 5 and Fable 5 at medium effort.
For computer use on OSWorld 2.0, GPT-6 Sol at xhigh effort reaches 60.5%, matching Claude Opus 5 at medium effort at roughly 80% lower cost.
Improved Collaboration Style
Both models inherit GPT-6 Astra’s communication improvements, offering clearer responses, less jargon, and more concise answers without losing substance. The models are now available in ChatGPT Work, Codex, and the OpenAI API.