OpenAI’s Strategic Price Cut For GPT‑6 Sol And Luna: Benchmark Scores Intact
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: OpenAI’s Strategic Price Cut For GPT‑6 Sol And Luna: Benchmark Scores Intact on ThorstenMeyerAI.com

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

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

TL;DR

OpenAI has announced a 50% price reduction for its GPT‑6 Sol and Luna models, with benchmark scores remaining stable. This shift aims to make AI more affordable for a wider range of applications, while performance metrics show mixed results.

OpenAI has announced a significant price cut for its GPT‑6 Sol and Luna models, reducing costs by approximately 50% while maintaining their benchmark performance levels. This move aims to broaden the accessibility of advanced AI models for businesses and developers, emphasizing cost-efficiency over new capabilities. The announcement comes just two weeks after the release of GPT‑6 Astra, marking a strategic shift toward more affordable AI deployment.

On September 22, 2026, OpenAI revealed that its GPT‑6 Sol and Luna models are now priced at half of their previous GPT‑5.6 equivalents. The new pricing reflects improvements in caching and inference technologies, allowing OpenAI to lower operational costs and pass savings to customers. For example, 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 is priced at $0.10 for input and $0.50 for output tokens, compared to $0.20 and $1.20 previously.

Independent analysis from Artificial Analysis confirmed that the cost per task has roughly halved, with the models’ benchmark scores remaining stable or improving slightly. GPT‑6 Sol scored 48 on the Artificial Analysis Intelligence Index, significantly above the median of 25 for comparable models, with a context window of 872,000 tokens. Luna scored 37, with a 1 million token window. Despite the lower prices, the models show mixed results in specific tasks, with some regressions in knowledge work benchmarks, attributed by analysts to reduced presentation quality in outputs.

OpenAI emphasized that the improvements are driven by technical enhancements, including more effective caching strategies, which reduce token read costs by 90%. The models also feature adjustable reasoning effort levels, allowing users to balance performance and cost. While the models’ core capabilities remain comparable to previous versions, some quality metrics, particularly in detailed knowledge work, have seen slight declines, according to independent testing.

At a glance
updateWhen: announced September 22, 2026
The developmentOpenAI has lowered the prices of GPT‑6 Sol and Luna models by half, maintaining their benchmark scores, to increase affordability and adoption.

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 Price Reduction on AI Adoption and Cost Efficiency

The price cuts for GPT‑6 Sol and Luna represent a strategic shift toward making advanced AI more accessible to a broader range of users and applications. By maintaining benchmark performance while halving costs, OpenAI enables smaller companies, startups, and research teams to incorporate powerful models into their workflows without prohibitive expenses. This could accelerate AI adoption across industries, fostering innovation and operational efficiencies. However, the mixed results in some quality metrics indicate that cost savings may come with trade-offs in output presentation and detailed knowledge tasks, which users should consider when integrating these models into critical workflows.

Amazon

AI development tools

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Background on OpenAI’s Model Pricing and Performance Trends

OpenAI’s release of GPT‑6 Astra two weeks prior marked the company’s entry into a new generation of models, emphasizing advanced capabilities. Historically, OpenAI has priced its models based on performance tiers, with higher-end models commanding premium prices. The recent announcement of GPT‑6 Sol and Luna at half the previous cost aligns with a broader industry trend of improving cost-efficiency through technical innovations, such as caching and inference optimizations. Independent evaluations, including those from Artificial Analysis, have shown that while costs can be reduced significantly, performance in certain knowledge-intensive tasks may not improve proportionally, highlighting ongoing challenges in balancing cost and quality.

Amazon

AI model deployment hardware

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Unresolved Questions About Long-Term Performance and Adoption

It remains unclear how these models will perform in large-scale, real-world deployments over extended periods, especially regarding the observed regressions in knowledge work benchmarks. The impact of reduced presentation quality on tasks requiring detailed, well-structured outputs is still being evaluated. Additionally, the long-term effects of lower costs on market competition and user adoption are uncertain, as other AI providers may respond with their own pricing strategies.

Amazon

AI token cost management software

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Next Steps for OpenAI and User Adoption Strategies

OpenAI is expected to continue refining its caching and inference techniques to further reduce operational costs and improve output quality. Users and developers should monitor upcoming updates and conduct thorough testing before fully deploying these models in critical workflows. Market responses, including potential competitive pricing from rivals like Anthropic, will influence how widely these models are adopted and integrated into various applications. OpenAI may also release further performance benchmarks or user case studies to demonstrate the models’ capabilities in different contexts.

Amazon

AI performance benchmarking tools

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Key Questions

What is the main benefit of OpenAI’s price cut for GPT‑6 Sol and Luna?

The primary benefit is a significant reduction in operational costs, enabling broader access and deployment of advanced AI models at half the previous price, while maintaining benchmark performance levels.

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

Independent analysis indicates some regressions in detailed knowledge work and presentation quality, though core capabilities and benchmark scores remain stable or improved. Users should evaluate these factors based on their specific use cases.

How do the new models compare in terms of speed and efficiency?

GPT‑6 Luna achieves 154 tokens per second at max effort, with a 1 million token context window, while Sol scores 115 tokens per second. Both models benefit from improved caching strategies, which reduce costs and improve response times in practical applications.

Will the price reduction impact OpenAI’s market position?

Lower prices could make OpenAI’s models more competitive, especially against rivals like Anthropic, which recently cut prices by 20%. The move may accelerate adoption but also prompts competitors to adjust their pricing strategies.

What should users consider before switching to these models?

Users should test the models thoroughly, especially for tasks requiring detailed, well-structured outputs, as some quality regressions have been noted. Balancing cost savings with output quality will be key in decision-making.

Source: ThorstenMeyerAI.com

EVERGREEN BESTSE

Evergreen bestsellers Picks

As an affiliate, we earn on qualifying purchases.

You May Also Like

CIA Funding Helped Keep NeXT Afloat In The 80S

Declassified documents reveal CIA funding helped keep NeXT afloat during the 1980s, raising questions about government involvement in tech startups.

Wireless Earbuds With Noise Cancellation: A Labor Day sales Guide

Discover how wireless earbuds with noise cancellation work, their pros and cons, and how to choose the right pair for your needs. Get the facts now.

Rocket Lab Surges In Global Coverage

Rocket Lab’s recent surge in international coverage, with 40 mentions in a short period, signals increased global interest in its space launch activities.

ByteDance AI: Why Its Strategic “Slow First, Fast Afterwards” Approach Is Reshaping The AI Industry – 36 Kr

ByteDance Seed describes its AI development approach as ‘slow first, fast afterwards,’ emphasizing early preparation before rapid deployment, but industry impact remains unverified.