OpenAI Unveils GPT-6 Sol and GPT-6 Luna, Cutting API Costs by 50% Across GPT-6 Frontier SeriesOpenAI has expanded its flagship model lineup with the simultaneous release of GPT-6 Sol and GPT-6 Luna. Arriving shortly after the debut of GPT-6 Astra, these two new additions complete the GPT-6 model tier, focusing heavily on accelerated inference throughput, ultra-low latency execution, and a dramatic reduction in enterprise API pricing.
Model Architecture, Benchmark Parity, and Price Disruption
The expanded GPT-6 series aligns high-throughput performance with disruptive operational cost metrics for enterprise developers:
Shared Architecture and High-Speed Throughput: Both GPT-6 Sol (the high-capacity performance variant) and GPT-6 Luna (the low-cost, high-speed variant) were trained using the same architectural methodologies as GPT-6 Astra. While matching core intelligence thresholds, Sol and Luna deliver significantly faster output generation speeds across complex workflows.
Frontier Benchmark Results: Standardized evaluations including AutomationBench, Agents’ Last Exam, and FrontierCode show GPT-6 Sol performing on par with or surpassing top-tier rival systems like the Claude series across autonomous coding, multi-step agent reasoning, and enterprise problem-solving tasks.
50% API Cost Reduction: OpenAI has slashed operational inference costs by 50% compared to previous-generation GPT-5.6 pricing. API rates for GPT-6 Sol are set at $2.00 per million input tokens and $10.00 per million output tokens, while GPT-6 Luna offers ultra-low rates of $0.10 per million input tokens and $0.50 per million output tokens.
Multi-Tier Ecosystem Rollout: Both models are immediately accessible through ChatGPT Work and Codex for Plus, Pro, Business, Enterprise, and Edu tier subscribers. Free and Go plan users receive native access to GPT-6 Luna via the desktop application, with web chat selector integration rolling out sequentially.
Source:
OpenAI
OpenAI Unveils GPT-6 Sol and GPT-6 Luna, Cutting API Costs by 50% Across GPT-6 Frontier SeriesOpenAI has expanded its flagship model lineup with the simultaneous release of GPT-6 Sol and GPT-6 Luna. Arriving shortly after the debut of GPT-6 Astra, these two new additions complete the GPT-6 model tier, focusing heavily on accelerated inference throughput, ultra-low latency execution, and a dramatic reduction in enterprise API pricing.
Model Architecture, Benchmark Parity, and Price Disruption
The expanded GPT-6 series aligns high-throughput performance with disruptive operational cost metrics for enterprise developers:
Shared Architecture and High-Speed Throughput: Both GPT-6 Sol (the high-capacity performance variant) and GPT-6 Luna (the low-cost, high-speed variant) were trained using the same architectural methodologies as GPT-6 Astra. While matching core intelligence thresholds, Sol and Luna deliver significantly faster output generation speeds across complex workflows.
Frontier Benchmark Results: Standardized evaluations including AutomationBench, Agents’ Last Exam, and FrontierCode show GPT-6 Sol performing on par with or surpassing top-tier rival systems like the Claude series across autonomous coding, multi-step agent reasoning, and enterprise problem-solving tasks.
50% API Cost Reduction: OpenAI has slashed operational inference costs by 50% compared to previous-generation GPT-5.6 pricing. API rates for GPT-6 Sol are set at $2.00 per million input tokens and $10.00 per million output tokens, while GPT-6 Luna offers ultra-low rates of $0.10 per million input tokens and $0.50 per million output tokens.
Multi-Tier Ecosystem Rollout: Both models are immediately accessible through ChatGPT Work and Codex for Plus, Pro, Business, Enterprise, and Edu tier subscribers. Free and Go plan users receive native access to GPT-6 Luna via the desktop application, with web chat selector integration rolling out sequentially.
Source:
OpenAI
Comments
Post a Comment