🔍 Read the full analysis: GPT‑6 Sol And Luna Prices Cut In Half, Benchmark Scores Remain Steady on ThorstenMeyerAI.com
Get hardware and tech essentials delivered free — and shop member deals
- Fast, free delivery on millions of items
- Access to Prime Big Deal Days deals on October 6–7
- Prime Video, Amazon Music and more included
TL;DR
OpenAI has reduced the prices of GPT-6 Sol and Luna models by half, with no significant change in their benchmark scores. Cost savings could expand AI adoption in various workflows, though some quality regressions are noted.
OpenAI has sharply reduced the prices of its GPT-6 Sol and GPT-6 Luna models by approximately 50%, making them significantly more affordable for businesses and developers. The models’ benchmark scores, including measures of intelligence and coding performance, have remained steady despite the price cuts. This development could influence how organizations incorporate AI into their workflows, expanding access to advanced language models at a lower cost.
On September 22, 2026, OpenAI announced that the prices for its GPT-6 Sol and Luna models have been halved compared to their GPT‑5.6 predecessors. GPT‑6 Sol now costs $2.00 per 1 million input tokens and $10.00 per 1 million output tokens, down from $4 and $20 respectively. Similarly, GPT‑6 Luna’s prices are now $0.10 and $0.50, compared to $0.20 and $1.20 previously. The company attributes these reductions to improvements in caching and inference techniques, which lower operational costs and allow passing savings to users.
Independent analysis from Artificial Analysis confirms that while costs per task have dropped roughly 50-60%, the models’ benchmark scores have remained stable or improved slightly in some evaluations. 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 of 12, with a 1 million-token context window. Despite the price cuts, the models’ core capabilities—such as reasoning, coding, and hallucination reduction—are largely unchanged, though some regressions in knowledge tasks have been observed.
OpenAI emphasizes that the cost reductions come without sacrificing the models’ core performance, which could make AI deployment more economically viable for a broader range of applications, from customer support to content generation. However, some trade-offs are noted, including a slight increase in hallucination rates and a reduction in presentation quality for some knowledge tasks.
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 Lower Costs on AI Adoption and Business Strategies
The 50% price reduction for GPT‑6 Sol and Luna models significantly lowers the entry barrier for organizations seeking to incorporate advanced AI into their operations. This could accelerate AI adoption across industries, especially for tasks where cost constraints previously limited deployment. The stable benchmark scores suggest that this affordability does not come at the expense of core model capabilities, potentially broadening AI’s practical reach. Nonetheless, the noted regressions in some knowledge and presentation tasks highlight the importance of testing models within specific workflows before full integration, as quality trade-offs may affect certain use cases.
AI language model API key management tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Background on GPT‑6 Model Pricing and Performance
OpenAI introduced GPT‑6 Astra earlier in September 2026, emphasizing improvements in cost efficiency and performance. The new models, Sol and Luna, are positioned as more affordable options within the GPT‑6 family, designed to democratize access to powerful AI. Prior to this, GPT‑5.6 models were the standard, with higher costs limiting widespread use. The recent price cuts follow technical enhancements in caching and inference, which OpenAI claims enable lower operational costs and pass those savings to users. Independent evaluations, such as those from Artificial Analysis, have shown that despite the lower prices, the models maintain competitive benchmark scores, though some regressions in knowledge tasks have been reported.
This move aligns with OpenAI’s broader strategy to expand AI deployment by making models more accessible, especially in workflows where cost efficiency is critical. The models’ performance in various benchmarks remains comparable to previous versions, with some noted exceptions in specific knowledge and presentation tasks, which may influence their suitability for certain applications.
As an affiliate, we earn on qualifying purchases.
Remaining Questions About Model Performance and Use Cases
It is still unclear how these models will perform in real-world, high-stakes applications, especially given some regressions observed in knowledge tasks and presentation quality. The long-term impact of reduced hallucination rates and the models’ ability to handle complex reasoning in diverse scenarios remains to be seen. Additionally, the full extent of cost savings in large-scale deployments and how organizations will adapt their workflows to leverage these lower prices are still developing topics.
cost-effective AI chatbot development kits
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Future Developments and Evaluation of GPT‑6 Models in Practice
OpenAI is expected to continue refining GPT‑6 Sol and Luna, with ongoing updates aimed at improving knowledge accuracy and presentation quality. Organizations are advised to conduct thorough testing within their specific workflows before full deployment, especially for tasks requiring high fidelity and detailed outputs. Further independent evaluations and user feedback will clarify how well the models perform at scale and whether the cost reductions translate into broader AI adoption across industries.
AI model performance monitoring tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
Why did OpenAI reduce the prices of GPT‑6 Sol and Luna models?
OpenAI achieved cost savings through technical improvements in caching and inference, which lower operational costs and enable passing those savings to users.
Do the benchmark scores indicate the models are less capable?
No, independent analysis shows that benchmark scores remain stable or slightly improve, suggesting core capabilities are maintained despite the price cuts.
Are there any drawbacks to the lower-priced models?
Some regressions in knowledge tasks and presentation quality have been observed, and increased hallucination rates may affect certain use cases requiring high accuracy.
How might these price reductions affect AI adoption?
The significant cost savings could make advanced AI more accessible for a broader range of organizations, potentially accelerating deployment in various industries.
What should organizations do before fully adopting these models?
Organizations should conduct thorough testing within their workflows to ensure the models meet their quality and performance standards, especially for critical tasks.
Source: ThorstenMeyerAI.com
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
