🔍 Read the full analysis: How OpenAI’s Price Cuts For GPT‑6 Sol And Luna Preserve Benchmark Stability on ThorstenMeyerAI.com
Get the latest gadgets 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 announced a 50% price reduction for its GPT‑6 Sol and Luna models, achieved through improved caching and inference. Despite lower costs, benchmark performance remains stable, though some regressions are noted in certain knowledge tasks. This shift could significantly influence AI adoption and operational costs.
OpenAI has announced a 50% reduction in the prices for its GPT‑6 Sol and Luna models, effective immediately, while maintaining stable benchmark performance. This move aims to make AI more accessible and cost-effective for a broad range of applications, from customer service to research support, with the potential to significantly lower operational expenses for businesses integrating these models.
The pricing for GPT‑6 Sol now stands at $2.00 per 1 million input tokens and $10.00 per 1 million output tokens, down from $4 and $20 respectively. Luna’s prices are halved to $0.10 and $0.50 per million tokens, previously $0.20 and $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 to customers.
Independent analysis by Artificial Analysis confirms that the cost per task has approximately halved, with costs dropping from about $1.99 to $1.06 for Sol and from $0.20 to $0.07 for Luna at maximum effort. Despite the lower prices, benchmark scores on intelligence and coding tasks remain strong, with Sol scoring 48 and Luna 37 on the Artificial Analysis Intelligence Index, well above median scores for models in their price class.
However, some regressions have been observed in knowledge-based evaluations, with Sol and Luna losing Elo points on certain economic and knowledge work benchmarks. These issues are linked to reduced presentation quality and omitted rubric elements, as noted by Artificial Analysis after manual review. OpenAI’s own release notes mention a focus on shorter, less detailed responses, which may impact workflows requiring comprehensive outputs.
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 Price Cuts on AI Cost and Adoption
The 50% price reduction for GPT‑6 Sol and Luna models represents a significant shift in the AI landscape, making large language models more affordable for a wider range of users and applications. By lowering operational costs, this move could accelerate AI adoption across industries, especially for tasks where cost-efficiency is critical. It also underscores OpenAI’s focus on broadening access to AI technology, potentially increasing competition and innovation in the sector.
While performance benchmarks remain stable overall, the noted regressions in some knowledge tasks highlight the importance of testing models within specific workflows before deployment. The trade-off between shorter responses and detailed deliverables could influence how businesses choose to integrate these models into their operations.
AI development and deployment tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Background on OpenAI’s Pricing and Model Developments
OpenAI’s GPT‑6 Astra was released two weeks prior, marking a new high in AI capabilities. The company’s strategy emphasizes not just improving intelligence but also reducing costs, enabling more extensive deployment. Prior to this, GPT‑5.6 models were priced higher, limiting their use to high-budget applications.
The recent price cuts for Sol and Luna, announced on September 22, 2026, follow OpenAI’s technical improvements in caching and inference, which have allowed a substantial decrease in costs while maintaining comparable benchmark scores. This approach aligns with industry trends toward making AI more accessible and affordable, especially as models become more integrated into everyday business workflows.
Independent evaluations, such as those from Artificial Analysis, provide detailed insights into the performance and cost-effectiveness of these models, confirming that the price reductions do not come at the expense of overall quality, although some specific knowledge tasks may see regressions.
cost-effective AI model hosting services
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Unresolved Questions About Long-Term Performance
It is not yet clear how these models will perform in real-world, long-term deployments, especially in complex knowledge tasks or specialized domains. The observed regressions in some benchmarks suggest potential limitations in certain workflows, and further testing is needed to determine if these issues persist over time or across different use cases.
Additionally, the impact of shorter, more concise responses on user satisfaction and accuracy remains to be fully evaluated in operational environments.
AI model caching and inference hardware
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps for Adoption and Model Evaluation
OpenAI is expected to continue refining these models, possibly addressing the identified regressions through future updates. Organizations planning to adopt GPT‑6 Sol and Luna should conduct thorough testing within their specific workflows, especially for tasks requiring detailed knowledge or comprehensive outputs.
Further independent evaluations and real-world case studies will clarify the models’ performance over extended periods, guiding strategic decisions around deployment and integration.
AI performance benchmarking software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
How much are GPT‑6 Sol and Luna models now priced?
GPT‑6 Sol costs $2.00 per 1 million input tokens and $10.00 per 1 million output tokens. Luna costs $0.10 and $0.50 respectively, representing a 50% reduction from previous prices.
Do the models still perform well on benchmark tests?
Yes, both models maintain strong benchmark scores, with Sol scoring 48 and Luna 37 on the Artificial Analysis Intelligence Index, comparable to previous versions.
Are there any downsides to the new pricing?
Some knowledge benchmarks have shown regressions, possibly due to changes aimed at shorter responses, which could affect workflows needing detailed outputs.
What technical improvements enabled the price cuts?
Enhanced caching and inference techniques have reduced operational costs, allowing OpenAI to lower prices while maintaining performance.
When will OpenAI release further updates?
OpenAI has not specified exact timelines, but ongoing model improvements and evaluations are expected as part of their development cycle.
Source: ThorstenMeyerAI.com
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
