🔍 Read the full analysis: GPT‑6 Sol And Luna Price Drops: OpenAI Maintains Benchmark Stability on ThorstenMeyerAI.com
Get everyday 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 announced price reductions for GPT‑6 Sol and Luna models, cutting costs by half compared to GPT‑5.6. Despite lower prices, performance remains stable overall, though some regressions in knowledge tasks are noted. The move aims to expand AI adoption through improved affordability.
OpenAI has introduced reduced-price versions of its GPT‑6 models, GPT‑6 Sol and Luna, with prices halved compared to GPT‑5.6. The models, launched on September 22, 2026, are designed to make advanced AI more accessible for a broader range of applications while maintaining comparable performance levels, according to the company.
Both models benefit from improved caching and inference techniques, enabling OpenAI to offer them at significantly lower costs. GPT‑6 Sol now costs $2.00 per 1 million tokens for input and $10.00 for output, down from $4 and $20 respectively, while GPT‑6 Luna costs $0.10 and $0.50, down from $0.20 and $1.20. These reductions reflect a 50% decrease against GPT‑5.6 promotional pricing, with the company passing savings on to customers.
Independent analysis by Artificial Analysis confirms that the cost per task has roughly halved, with minimal impact on overall AI scores. GPT‑6 Sol’s Intelligence Index score remains high at 48, well above the median of 25 for models in its class, with Luna scoring 37 against a median of 12. Performance in coding tasks also improved slightly, with Sol scoring 57 on the Coding Agent Index, and Luna scoring 41.
However, some regressions are noted in knowledge-related tasks, with GPT‑6 Sol dropping approximately 100 Elo points on certain economic benchmarks, and Luna falling about 75 points. These regressions are attributed to changes in output presentation quality and omitted details, as per the analysis.
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 of Lower-Cost GPT‑6 Models for AI Adoption
This price reduction makes advanced AI more affordable for a wider range of businesses and developers, potentially accelerating adoption across industries. By maintaining performance levels while cutting costs, OpenAI enhances the viability of deploying large language models in operational workflows, customer service, and content generation. However, the noted regressions in knowledge tasks suggest that users should test these models within their specific contexts to ensure quality standards are met.
AI language model API subscription
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’s GPT‑6 family was launched in September 2026, following the Astra model’s release two weeks earlier. The Astra model set a new standard for AI capabilities, but the real innovation lies in making these capabilities accessible through significant price reductions. Prior to this, GPT‑5.6 models were priced higher, limiting their use in cost-sensitive applications. The new models leverage advancements in caching and inference, allowing OpenAI to lower operational costs and pass savings to users.
Independent evaluations by Artificial Analysis have shown that while costs have decreased substantially, performance metrics remain largely stable, with some areas experiencing regressions. The focus on cost efficiency aims to democratize AI deployment, especially for tasks where budget constraints previously limited adoption.
cost-effective AI coding assistant
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Remaining Questions About Model Performance and Adoption
It is still unclear how these models will perform across diverse real-world applications over longer periods. The observed regressions in knowledge tasks raise questions about their suitability for detailed research or complex decision-making workflows. Additionally, the long-term impact of the reduced presentation quality on user satisfaction remains to be seen, and OpenAI has not yet provided detailed user feedback or case studies.
As an affiliate, we earn on qualifying purchases.
Next Steps for OpenAI and Users of GPT‑6 Models
OpenAI is expected to continue refining its models and caching techniques, potentially addressing the regressions in knowledge tasks. Users are advised to conduct thorough testing within their specific workflows before full deployment. Industry analysts anticipate that OpenAI may release further enhancements or new versions in the coming months, further optimizing cost and performance. Monitoring user feedback and independent evaluations will be critical to assess long-term viability.
affordable AI chatbot for business
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 approximately $2.00 per 1 million tokens for input and $10.00 for output, a 50% reduction from GPT‑5.6. Luna costs $0.10 and $0.50, respectively, about 60% less than its predecessor.
Do the new models perform worse than previous versions?
Overall, performance remains stable or improved in many areas, as confirmed by independent analysis. However, some knowledge and reasoning tasks show regressions, likely due to changes in output presentation and detail omission.
What are the main improvements in GPT‑6 Sol and Luna?
Major improvements include lower costs, better caching for prompt reuse, and slight gains in coding and reasoning scores. Hallucination rates have also decreased significantly.
Should I switch to these models now?
It depends on your use case. For cost-sensitive applications, testing is recommended to ensure output quality meets your needs, especially in knowledge-intensive tasks.
What is likely to happen next in OpenAI’s model development?
OpenAI will likely continue refining these models, addressing current regressions, and possibly releasing newer versions with further performance improvements and cost reductions.
Source: ThorstenMeyerAI.com
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
