Understanding OpenAI’s Price Drop For GPT‑6 Sol And Luna Amid Stable Benchmarks
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: Understanding OpenAI’s Price Drop For GPT‑6 Sol And Luna Amid Stable Benchmarks on ThorstenMeyerAI.com

Prime Big Deal Days · Oct 6–7Offer from Amazon

Get smart everyday buys 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
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

TL;DR

OpenAI announced a 50% price reduction for GPT‑6 Sol and Luna models, effective September 22, 2026, without significant changes in benchmark performance. The move aims to make AI more accessible for business use, with cost savings driven by improved caching and inference efficiencies.

OpenAI has significantly reduced the prices of its GPT‑6 Sol and GPT‑6 Luna models by 50% effective September 22, 2026. The models now cost half as much as their GPT‑5.6 predecessors, with no major changes in benchmark performance. This move aims to democratize access to advanced AI by lowering operational costs for businesses and developers, emphasizing cost efficiency over new capabilities.

OpenAI introduced GPT‑6 Sol and Luna on September 22, 2026, with prices halved compared to previous GPT‑5.6 models. 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. Luna’s prices are $0.10 and $0.50, compared to $0.20 and $1.20. These reductions are attributed to improvements in caching and inference technology, which allow OpenAI to serve these models at lower costs. According to independent analysis by Artificial Analysis, the models’ performance, measured by composite scores and task efficiency, remains roughly stable despite the price cuts. GPT‑6 Sol scores 48 on the Artificial Analysis Intelligence Index, well above the median of 25, with a 872,000-token context window, while Luna scores 37, surpassing the median of 12, with a 1 million-token window. Cost per task has halved, with Sol at maximum effort costing about $1.06 and Luna about $0.07, compared to their predecessors. The models show improvements in hallucination reduction, with Sol decreasing hallucination rates from 92% to 60%, and Luna from 93% to 77%. However, some regressions were noted in knowledge-work benchmarks, with both models scoring lower on certain economic and productivity evaluations, likely due to adjustments aimed at reducing low-value outputs. OpenAI emphasizes that caching enhancements, including 90% discounts on cached input reads, contribute significantly to the reduced costs, alongside inference efficiency gains. These developments reflect a strategic focus on making AI models more affordable for enterprise and developer use, broadening the scope of AI deployment across industries.

At a glance
updateWhen: announced September 22, 2026
The developmentOpenAI has lowered the prices of its GPT‑6 Sol and Luna models by half, while performance benchmarks remain stable, marking a strategic shift toward cost-effective AI deployment.

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.

GPT‑6 Sol
$4 / $20 → $2 / $10
GPT‑6 Luna
$0.20 / $1.20 → $0.10 / $0.50

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.

GPT‑5.6 Sol
$1.99
GPT‑6 Sol
$1.06
GPT‑5.6 Luna
$0.18
GPT‑6 Luna
$0.07

The effort dial moves cost more than the model choice

Model and effortIntelligence IndexCost 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

Already on GPT‑5.6 Sol or Luna? The move is mostly a price cut. Re‑test first if your output is a document someone reads, not data a system consumes.
Shelved an automation on cost? Token prices halved and the effort dial adds another order of magnitude. Re‑run the business case.
Choosing between labs? The question is no longer which model is smartest, but which clears your quality bar at the lowest cost per task.
ThorstenMeyerAI.comSources: OpenAI (pricing, vendor benchmarks) and Artificial Analysis (independent evaluation and model pages). Figures as of 23 September 2026.

Implications of Lower AI Model Costs

The price reductions for GPT‑6 Sol and Luna are poised to broaden AI adoption by significantly lowering operational costs for businesses and developers. This shift enables more organizations to incorporate advanced AI into their workflows without the previously prohibitive expenses, potentially accelerating AI-driven automation and innovation. While performance benchmarks remain stable, the improvements in cost efficiency could lead to increased deployment in customer service, research, and content generation. However, some quality regressions in specific knowledge tasks suggest that users should carefully evaluate models for their particular use cases, especially those requiring detailed, well-structured outputs. Overall, this move signals OpenAI’s strategic emphasis on democratizing AI access and fostering widespread adoption through affordability, which could reshape competitive dynamics in the AI industry.

Amazon

AI model cost management tools

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Background on OpenAI’s Pricing and Performance Trends

Prior to this announcement, OpenAI’s GPT‑6 models were priced at levels comparable to or higher than previous iterations, reflecting their advanced capabilities. The release of GPT‑6 Astra earlier in 2026 marked a milestone in AI intelligence, but the real shift came with the new Sol and Luna models, which focus on cost efficiency. Independent evaluations by Artificial Analysis, published concurrently, indicated that while costs per task have halved, the models’ performance on various benchmarks has remained stable or improved slightly, with some noted regressions in knowledge-oriented tasks. OpenAI’s ongoing improvements in caching and inference technology have been central to these cost reductions. Historically, OpenAI has balanced performance with cost considerations, but this latest move underscores a strategic pivot toward making AI more accessible and affordable for a broader range of applications.

Amazon

enterprise AI deployment software

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Unresolved Questions About Model Performance

While initial evaluations show stable performance, some regressions in knowledge-work benchmarks suggest that the models may have trade-offs in presentation quality and detailed output. It remains unclear how these regressions will impact long-term or specialized use cases, and whether further tuning will address these issues. Additionally, the full impact of caching improvements on operational efficiency across diverse deployment scenarios is still being assessed, and the models’ performance in real-world, high-demand environments remains to be seen.

Amazon

AI inference optimization hardware

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 refining its models and caching technology, possibly releasing updates that address current regressions. Businesses and developers are advised to test these models within their specific workflows to gauge performance and output quality. Monitoring user feedback and independent evaluations over the coming months will be critical to understanding the full impact of these price reductions. OpenAI may also expand its enterprise offerings or introduce new features to further support broader adoption of GPT‑6 models.

Amazon

AI caching and inference technology products

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?

OpenAI lowered prices due to technological improvements in caching and inference, which reduce operational costs while maintaining performance levels.

How do the new GPT‑6 models compare to previous versions in terms of performance?

Performance benchmarks remain roughly stable, with some improvements in hallucination reduction and efficiency, though some knowledge-work metrics show regressions.

What are the main benefits of the price cuts for businesses?

The lower costs enable wider adoption of advanced AI, making it more feasible for organizations to integrate AI into their workflows and products.

Are there any downsides or risks associated with these models?

Some regressions in detailed knowledge tasks suggest that model outputs may be less comprehensive or polished in certain contexts, warranting careful testing before deployment.

What should users expect next from OpenAI regarding these models?

OpenAI is likely to release further updates to improve performance and address current regressions, while expanding caching and inference efficiencies.

Source: ThorstenMeyerAI.com

This content is for general information only and is not financial, tax or legal advice. Consult a qualified professional for decisions about your money.
FALL

Fall Picks

As an affiliate, we earn on qualifying purchases.

You May Also Like

Aleph Alpha. The retrospective case.

Analyzing Aleph Alpha’s strategic pivot, founder departure, and recent merger with Cohere to understand the costs of late structural adaptation in European AI.

Using Management Tests To Understand AI’s Authentic Behavior

Exploring how management-style tests reveal true AI capabilities in decision-making, trust, and execution through live business simulations.

Streamline Tech Operations With A 500-Line C++ Signal Monitor

A new 500-line C++ signal monitor helps small software teams track platform and tooling changes efficiently, improving decision-making speed.

Airworthy.com Releases 2026 State Of General Aviation Parts Report

Airworthy.com has published its 2026 State of General Aviation Parts Report, highlighting industry trends and supply chain insights for the upcoming year.