🔍 Read the full analysis: Why OpenAI’s GPT‑6 Sol And Luna Are Now More Affordable While Benchmarks Stay Flat on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
OpenAI has launched GPT-6 Sol and Luna models at half the previous prices, with benchmarks remaining steady. This shift makes AI more accessible for a wider range of applications, though some quality regressions are noted.
OpenAI has introduced GPT‑6 Sol and GPT‑6 Luna models at approximately 50% lower prices than their GPT‑5.6 predecessors, with the company emphasizing cost efficiency over new capabilities. This move aims to expand AI accessibility for a broader range of applications, as the models’ benchmarks remain largely stable.
Both models, GPT‑6 Sol and Luna, were launched on September 22, 2026, with significant reductions in cost: GPT‑6 Sol now costs $2.00 per 1 million input tokens (down from $4), and $10.00 per 1 million output tokens (down from $20). GPT‑6 Luna costs $0.10 per 1 million input tokens (down from $0.20) and $0.50 per 1 million output tokens (down from $1.20). These reductions are attributed to improvements in caching and inference techniques, which allow OpenAI to serve these models at lower costs, passing savings onto customers.
Independent analysis by Artificial Analysis, published concurrently, confirms that while the costs per task have halved, the models’ performance benchmarks remain roughly steady. GPT‑6 Sol scores 48 on the Artificial Analysis Intelligence Index, well above the median of 25, with a 872,000-token context window. Luna scores 37, also above the median, with a 1-million-token context window. Despite the lower costs, some evaluations, especially in knowledge work, show regressions, which OpenAI attributes to changes in presentation quality and output formatting.
GPT‑6 Sol and Luna: half the price, about the same intelligence
OpenAI’s September 22, 2026 release doesn’t raise the ceiling. It lowers the cost of everything below it, which changes what’s worth automating.
Per 1M input / output tokens. Cached input reads keep the 90% discount.
Cost per task, halved
Measured by Artificial Analysis as the weighted cost of one Intelligence Index task, at max effort.
The effort dial moves cost more than the model choice
| Model and effort | Intelligence Index | Cost per task |
|---|---|---|
| GPT‑6 Sol (max) | 48 | $1.06 |
| GPT‑6 Sol (low) | 34 | $0.13 |
| GPT‑6 Luna (max) | 37 | $0.07 |
| GPT‑6 Luna (low) | 21 | $0.0045 |
| GPT‑6 Luna (non‑reasoning) | 18 | $0.01 |
Sol at low effort keeps about 70% of its max score for roughly an eighth of the cost, because it writes far fewer reasoning tokens. For reference, Claude Opus 5.5 leads the same index at 58.
What got better, and what got worse
Better
- Hallucination rate on AA‑Omniscience: Sol 92% → 60%, Luna 93% → 77%
- Coding Agent Index: Sol 57, up 2 points, at ~50% lower cost per task
- OpenAI reports about half as many factual mistakes for Sol as its predecessor
- Higher cache hit rates; GitHub reports over 50% fewer prompt tokens needing fresh processing
Sol gets there partly by declining more: it attempts 83% of questions vs 99%, and accuracy falls 59% → 54%.
Worse
- GDPval‑AA v2.1: Sol down ~100 Elo, Luna down ~75
- AA‑Briefcase v1.1: Luna down ~45 Elo
- Coding Agent Index: Luna 41, down 2 points
- Both models write more output tokens per task than their predecessors
Reviewers attribute the drops to weaker presentation and deliverables that omit required elements.
What to do about it
Implications for AI Deployment and Cost Management
The reduction in model prices significantly broadens the potential for AI integration across industries, enabling smaller firms and projects to leverage advanced language models without prohibitive costs. This shift could accelerate automation, research, and customer service applications, making AI-driven solutions more feasible and widespread. However, the noted regressions in some performance metrics highlight the importance of testing models for specific workflows, especially where detailed, well-structured outputs are required.
As an affiliate, we earn on qualifying purchases.
Pricing and Performance Trends in AI Models
OpenAI’s move follows a broader industry trend of reducing AI model costs while maintaining performance benchmarks. The launch of GPT‑6 Sol and Luna at half-price reflects ongoing efforts to improve inference efficiency through caching and technical optimizations. Previously, AI models like Anthropic’s Claude Opus 5.5 also announced price cuts, indicating a competitive environment focused on balancing cost and capability. Despite these advances, some evaluations reveal performance regressions in knowledge and document production tasks, attributed to tuning for user experience rather than raw output quality.
affordable AI chatbot development tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Remaining Questions on Model Quality and Use Cases
It is not yet clear how the performance regressions in knowledge work and detailed document generation will impact real-world applications. The models’ lower hallucination rates are promising, but the trade-off with output completeness needs further testing across diverse workflows. Additionally, the long-term effects of the tuning changes on model robustness remain uncertain.
large language model tokens management
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps for Adoption and Evaluation
Organizations and developers are expected to begin integrating GPT‑6 Sol and Luna into their workflows, with ongoing testing to assess suitability for specific tasks. OpenAI will likely release further updates on model performance, caching tools, and cost management strategies. Monitoring user feedback and independent evaluations will be crucial to understanding the full impact of these models on AI deployment.
AI performance benchmarking software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
How much cheaper are GPT‑6 Sol and Luna compared to previous models?
GPT‑6 Sol costs about half as much per 1 million tokens as GPT‑5.6, with input at $2.00 and output at $10.00. Luna costs roughly 50-60% less, at $0.10 and $0.50 per million tokens, respectively.
Do these new models perform as well as their predecessors?
In most benchmarks, GPT‑6 Sol and Luna maintain similar performance levels, with some regressions noted in knowledge and document production tasks. They also show improved hallucination rates and cost efficiency.
What are the main trade-offs with these models?
While costs are reduced, some evaluations indicate a decline in presentation quality and detailed output, especially in knowledge work. Additionally, models may refuse to answer more often, which can affect workflows requiring comprehensive responses.
How might this price reduction affect AI adoption?
The lower prices make advanced AI models accessible to smaller organizations and a broader range of applications, potentially accelerating automation and AI-driven innovation across industries.
What should users do before switching to these models?
Users should conduct testing specific to their workflows, especially if detailed, high-quality outputs are critical, as some regressions in presentation and completeness have been observed.
Source: ThorstenMeyerAI.com
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
