AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: OpenAI Halves GPT‑6 Sol And Luna Costs With No Impact On Benchmarks on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on monitors, keyboards and dev gear

  • Business-only prices and quantity discounts
  • Tax-exempt purchasing
  • Multiple users, one account, clear invoices
As an affiliate, we earn on qualifying purchases.

TL;DR

OpenAI has halved the prices of its GPT‑6 Sol and Luna models without impacting their benchmark scores. The move aims to make AI more accessible for commercial use, with cost savings passed directly to users.

OpenAI has significantly reduced the costs of its GPT‑6 Sol and Luna models by 50%, with no loss in benchmark performance, marking a major shift toward making advanced AI more affordable for businesses and developers.

On September 22, 2026, OpenAI introduced a price cut for its GPT‑6 Sol and Luna models, decreasing costs for input and output tokens by half compared to previous GPT‑5.6 versions. The new pricing reflects improvements in caching and inference technologies, allowing the models to be served at lower costs, with the savings passed directly to customers. Specifically, 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 costs are halved to $0.10 and $0.50 per million tokens.

Independent analysis from Artificial Analysis confirms that while costs have decreased by about 50%, the models’ benchmark scores remain roughly stable. GPT‑6 Sol scores 48 on the Artificial Analysis Intelligence Index, significantly above its median, and Luna scores 37, also well above the median for its class. Despite the price cuts, the models’ performance in tasks such as coding and hallucination reduction has seen mixed results, with notable improvements in hallucination rates but some regressions in knowledge-based evaluations.

At a glance
updateWhen: announced September 22, 2026
The developmentOpenAI announced on September 22, 2026, that it has reduced the prices of GPT‑6 Sol and Luna models by 50%, while maintaining their performance benchmarks.

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.

Impact of Cost Reduction on AI Adoption

The price reductions for GPT‑6 Sol and Luna lower the barrier for integrating advanced AI into a wider range of applications, from customer service to automation workflows. By maintaining benchmark performance, OpenAI enables businesses to deploy these models without sacrificing quality, potentially accelerating AI adoption across industries. This shift could lead to more cost-effective AI solutions, expanding the scope of tasks that can be automated and improving operational efficiency for many organizations.

Amazon

AI model cost management tools

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Previous Pricing and Technological Improvements

Prior to this announcement, OpenAI’s GPT‑6 models were priced at levels comparable to or higher than competitors, limiting their accessibility for smaller firms or projects with tight budgets. The company attributes the new pricing to advancements in caching and inference, which have reduced computational costs. OpenAI’s release of Astra, its top-tier model, was aimed at high-end applications, while the new Sol and Luna models target more cost-sensitive markets. The move reflects a broader industry trend toward balancing performance with affordability, especially as AI becomes more embedded in commercial workflows.

Amazon

cloud-based AI inference hardware

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Unclear Long-term Effects on Model Performance

While initial benchmarks show stable performance, it is still unclear whether the cost reductions will sustain the models’ effectiveness in more complex, real-world scenarios over time. The impact on other evaluation metrics, such as long-term knowledge retention and nuanced reasoning, remains to be seen. Additionally, the effects of reduced presentation quality noted in some evaluations could influence certain use cases, especially those requiring detailed outputs.

Amazon

AI developer token pricing tools

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Next Steps for OpenAI and Model Adoption

OpenAI is expected to continue refining its caching and inference techniques to further reduce costs and improve efficiency. Monitoring how these models perform in diverse applications will be critical, as organizations test their capabilities in real-world environments. Additionally, OpenAI may release updated benchmarks and user feedback to validate long-term performance stability. Market adoption will likely increase as more developers and companies leverage these more affordable models for a broad array of AI-driven tasks.

Amazon

affordable AI model deployment hardware

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Key Questions

Will the quality of GPT‑6 Sol and Luna decline with the price cut?

According to OpenAI and independent evaluations, benchmark scores remain stable, but some quality aspects like presentation and detailed output may be slightly affected, especially in complex tasks.

How much will businesses save with the new pricing?

Cost savings are approximately 50% compared to previous GPT‑5.6 models, reducing input token costs from $4 to $2 and output from $20 to $10 for Sol, and similarly for Luna.

Does this price cut affect all use cases?

While cost reductions benefit most applications, some tasks requiring detailed, well-presented outputs might experience slight regressions. Testing is recommended before large-scale deployment.

Are there any performance trade-offs with the new models?

Performance in hallucination reduction has improved, but some knowledge-based assessments show regressions, likely due to tuning for better conversational quality rather than factual accuracy.

What is the significance for AI developers and users?

The lower prices make advanced AI more accessible, enabling broader experimentation and deployment across industries, potentially accelerating AI-driven innovation and automation.

Source: ThorstenMeyerAI.com

FALL

Fall Picks

As an affiliate, we earn on qualifying purchases.

You May Also Like

Software engineering. The canonical case.

New data shows a 40% drop in junior developer hiring since 2022, with senior engineers increasingly augmented by AI. The sector reveals a bifurcated impact of AI on labor.

Arista Networks Surges In Global Coverage

Arista Networks experiences a surge in worldwide coverage, with 33 mentions in recent media monitoring, indicating increased industry and public interest.

Forezai · Polybot: When the AI Disagrees With the Odds

Polybot, an AI trading tool, tests when an AI’s probability estimates diverge from prediction market prices, raising questions about market efficiency.

Understanding Anthropic’s $965B Series H: The Compute Revolution

Anthropic’s latest funding round signals a major shift toward infrastructure investments—chips, memory, and power—to support AI scaling at unprecedented levels.