🔍 Read the full analysis: GPT‑6 Sol And Luna Now Half The Price—AI Benchmarks Show No Change on ThorstenMeyerAI.com
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
TL;DR
OpenAI has launched GPT-6 Sol and Luna models at 50% lower prices than GPT‑5.6, with independent benchmarks indicating no major change in AI performance. The move aims to expand cost-effective AI deployment across industries.
OpenAI has introduced GPT-6 Sol and Luna models at half the previous prices, marking a significant shift in AI affordability without altering the models’ performance benchmarks. The models, released on September 22, 2026, are designed to make AI deployment more accessible for a broader range of applications, from customer service to research. This move underscores a focus on cost efficiency rather than pushing the boundaries of AI capabilities, which remains stable according to independent evaluations.
Both GPT-6 Sol and Luna are now priced at approximately 50% less than their GPT‑5.6 predecessors. The pricing reductions are achieved through improvements in caching and inference efficiency, allowing OpenAI to pass savings directly to users. GPT-6 Sol costs $2.00 per 1 million tokens for input and $10.00 for output, down from $4 and $20 respectively. Luna costs $0.10 for input and $0.50 for output, compared to $0.20 and $1.20 previously. Despite the lower prices, independent benchmarks from Artificial Analysis show no significant decline in the models’ AI performance scores.
Independent testing indicates that cost per task has roughly halved, with GPT-6 Sol at $1.06 per task and Luna at $0.07,
compared to their GPT‑5.6 counterparts. The models demonstrate stable or improved scores on key benchmarks like the Artificial Analysis Intelligence Index, with Sol scoring 48 and Luna 37—well above the median in their respective classes. However, some regressions are noted in knowledge-based assessments, attributed to changes in output presentation rather than core intelligence capabilities.
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
Impact of Cost Reduction on AI Deployment
The reduction in model prices by 50% significantly lowers the barrier to entry for businesses and developers seeking to incorporate AI into their workflows. This shift could accelerate adoption in sectors like customer support, content generation, and research, where budget constraints previously limited AI use. The unchanged performance benchmarks reassure users that cost savings do not come at the expense of quality, making these models attractive for scalable, cost-sensitive applications.
By enabling more affordable AI, OpenAI’s move could reshape the competitive landscape, putting pressure on other providers to follow suit. It also emphasizes a strategic focus on cost efficiency as a key driver of AI democratization, rather than solely advancing model capabilities. However, the unchanged benchmarks also highlight that the models’ core intelligence remains stable, and improvements are primarily in operational efficiency and pricing.
As an affiliate, we earn on qualifying purchases.
Background on GPT Model Pricing and Performance
Since the launch of GPT‑5.6, OpenAI has steadily worked on improving the cost efficiency of its models through technical innovations such as caching and inference optimization. The release of GPT‑6 Astra earlier this year marked a milestone in AI capabilities, but the recent focus is on making these capabilities more accessible financially. Historically, AI model pricing has been a barrier for widespread adoption, especially for smaller businesses and startups.
The introduction of GPT‑6 Sol and Luna follows industry trends toward more affordable AI, with competitors like Anthropic also reducing prices—Claude Opus 5.5, for example, cut costs by 20%. Prior to this, OpenAI’s models were considered state-of-the-art but costly, limiting their use to large enterprises or specialized applications. The new models aim to change that by offering similar performance at a drastically lower cost, broadening the scope of AI’s practical applications.
Independent evaluations, such as those from Artificial Analysis, have shown that while the models’ operational costs have decreased, their core intelligence and accuracy benchmarks have remained stable or improved slightly, confirming that the price cut does not equate to a performance downgrade.
cost-effective AI chatbot software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Unanswered Questions About Long-Term Performance
It is still unclear how the models will perform over extended use or in more complex, real-world scenarios. While initial benchmarks show stability, ongoing testing is required to verify whether the models maintain their performance in diverse applications. Additionally, the impact of reduced presentation quality and output detail—observed in some evaluations—may influence user satisfaction and effectiveness in production environments.
Further, the long-term effects of caching improvements on model responsiveness and cost savings are still being assessed. The models’ ability to handle more nuanced tasks or maintain accuracy over time remains an open question.
AI development tools for businesses
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps for Adoption and Evaluation
OpenAI is expected to continue monitoring the models’ performance in various deployment scenarios, gathering user feedback, and refining caching and inference techniques. Industry analysts anticipate that more organizations will begin integrating GPT‑6 Sol and Luna into their workflows, testing their limits in real-world tasks. OpenAI may also release further updates or new models aimed at optimizing both cost and performance.
Meanwhile, independent researchers and enterprise users will likely conduct extended evaluations, especially focusing on long-term stability, accuracy in complex tasks, and user satisfaction. The competitive landscape could also see adjustments as other AI providers respond to OpenAI’s aggressive pricing strategy.
affordable AI content generation software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
Are GPT‑6 Sol and Luna less capable than previous models?
No, independent benchmarks indicate that their performance remains stable or slightly improved in key areas, despite the lower prices.
What are the main benefits of the price reduction?
The primary benefit is making AI more affordable and accessible for a broader range of applications and organizations, reducing operational costs significantly.
Will the lower prices affect the models’ quality in real-world use?
Initial benchmarks suggest performance stability, but ongoing real-world testing will determine if quality remains consistent across diverse applications.
How do caching improvements contribute to cost savings?
Enhanced caching allows more reuse of context, reducing the computational load and enabling lower token prices without sacrificing output quality.
When will OpenAI release more models or updates?
OpenAI has not announced specific timelines, but continued performance monitoring and user feedback are expected to guide future releases.
Source: ThorstenMeyerAI.com
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
