🔍 Read the full analysis: OpenAI’s Price Cuts For GPT‑6 Sol And Luna: Maintaining Benchmark Performance on ThorstenMeyerAI.com
Get hardware and tech 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 has announced a 50% price reduction for its GPT‑6 Sol and Luna models, aiming to make AI more accessible without sacrificing performance. The models now offer significantly lower costs per task, with mixed results in accuracy and quality improvements. The development signals a focus on cost efficiency, but some regressions in knowledge tasks have been observed.
OpenAI has announced a significant reduction in the prices of its GPT‑6 Sol and Luna models, slashing costs by 50% compared to GPT‑5.6 predecessors. The models, introduced on September 22, 2026, are designed to make advanced AI more affordable for a broader range of applications, emphasizing cost efficiency over pushing new capability boundaries.
The new models, GPT‑6 Sol and Luna, are priced at $2.00 and $0.10 per 1 million tokens for input, and $10.00 and $0.50 per 1 million tokens for output, respectively. These prices reflect improvements in caching and inference technologies that reduce operational costs, with the savings passed directly to users.
Artificial Analysis’s evaluation confirms that costs per task have approximately halved, with GPT‑6 Sol costing about $1.06 per task and Luna around $0.07, both roughly 50-60% less than their GPT‑5.6 equivalents. Despite the lower costs, performance metrics show mixed results: Sol scores 48 on the Artificial Analysis Intelligence Index, significantly above the median, while Luna scores 37, also above average in its class.
Quality improvements include a notable reduction in hallucination rates—Sol’s hallucination rate dropped from 92% to 60%, and Luna’s from 93% to 77%. However, this came at the expense of increased refusals to answer, with Sol attempting fewer questions, which lowered its accuracy from 59% to 54%. Some knowledge benchmarks, such as GDPval‑AA v2.1, showed regressions, attributed to reduced presentation quality and omitted details, indicating a trade-off between cost and completeness.
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 Models for AI Deployment
The price cuts for GPT‑6 Sol and Luna are poised to significantly expand AI adoption across industries, particularly in workflows where cost constraints previously limited use. By offering models that maintain high benchmark scores at half the previous prices, OpenAI is enabling organizations to automate more tasks, scale AI integration, and reduce operational costs. However, the observed regressions in some knowledge tasks highlight the importance of testing these models within specific workflows to ensure quality and reliability.
This shift could reshape the AI market by increasing competition and setting new cost-performance benchmarks, encouraging other providers to follow suit. For users, the development means more affordable access to powerful language models, but also underscores the need for careful evaluation of model outputs, especially in knowledge-intensive applications.
As an affiliate, we earn on qualifying purchases.
Background on OpenAI’s Pricing and Model Development
OpenAI’s recent model releases have focused on balancing performance with cost efficiency. The launch of GPT‑6 Astra earlier this year marked a new top-tier benchmark, but the introduction of Sol and Luna shifts emphasis toward affordability. The models are part of OpenAI’s broader strategy to democratize access to advanced AI by lowering entry barriers through technological improvements like caching and inference optimization.
Prior to this, GPT‑5.6 models were the standard, priced higher and used for high-stakes or high-quality tasks. The new models are designed to serve a wider range of applications, from customer service to content generation, where cost savings are critical. The evaluation by Artificial Analysis, published concurrently, confirms that while costs have halved, performance remains competitive, though some quality regressions have been noted in specific knowledge tasks.
affordable AI chatbot development tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Unresolved Questions About Model Long-Term Performance
It remains unclear how these models will perform in real-world, long-term deployments, especially in sensitive or high-stakes environments. The observed regressions in some knowledge benchmarks suggest that further tuning may be needed to optimize for specific use cases. Additionally, the impact of increased refusal rates on user experience and workflow efficiency is still being assessed, and OpenAI has not yet publicly detailed plans for ongoing updates or improvements.
GPT-6 compatible AI development software
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 GPT‑6 Sol and Luna based on user feedback and performance data. Industry analysts anticipate further performance tuning and possibly new versions that address current regressions. Meanwhile, organizations are advised to pilot these models carefully, testing their suitability for specific applications before fully integrating them into operational workflows. Monitoring of long-term performance and stability will be essential as adoption scales.
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’s costs are approximately 50% less, with input at $2.00 and output at $10.00 per million tokens. Luna’s costs are about 60% lower, at $0.10 for input and $0.50 for output per million tokens.
Do the new models perform as well as older versions?
Performance remains strong on benchmark scores, with Sol scoring 48 and Luna 37 on the Artificial Analysis Intelligence Index, both above median. However, some knowledge benchmarks show regressions, and quality in detailed output may be slightly reduced.
What are the main trade-offs with these new models?
The models achieve lower hallucination rates partly by refusing to answer more often, which can reduce completeness in some tasks. Regressions in knowledge tasks suggest a trade-off between cost, quality, and presentation detail.
Will OpenAI release further updates or new models?
OpenAI is likely to continue refining GPT‑6 Sol and Luna based on user feedback and performance data, possibly releasing improved versions to address current limitations.
How should organizations approach adopting these models?
Organizations should pilot these models carefully, testing their performance in specific workflows, especially for knowledge-intensive tasks, before full deployment.
Source: ThorstenMeyerAI.com
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
