🔍 Read the full analysis: OpenAI's Latest Move: Halving GPT‑6 Sol And Luna Prices Without Benchmark Disruption on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
OpenAI announced a 50% price reduction for GPT-6 Sol and Luna models, driven by improvements in caching and inference. Despite lower costs, benchmark scores remain stable overall, though some regressions are noted.
OpenAI has halved the prices of its GPT‑6 Sol and Luna models, effective immediately, aiming to make AI more accessible without sacrificing performance. The move, announced on September 22, 2026, is driven by technical improvements in caching and inference that reduce operational costs. This price reduction is significant because it broadens the potential applications of these models in commercial workflows and automation, where cost is a critical factor.
Both models are now offered at 50% lower prices than their GPT‑5.6 predecessors. GPT‑6 Sol’s input costs are now $2.00 per 1 million tokens, down from $4, while output costs are $10.00, down from $20. OpenAI attributes the savings to enhanced caching techniques that allow more efficient reuse of context, including a 90% discount on cached input reads. GPT‑6 Luna’s prices are $0.10 per 1 million input tokens and $0.50 per 1 million output tokens, halving previous costs of $0.20 and $1.20 respectively.
Independent analysis by Artificial Analysis confirms that the cost per task has roughly halved, with no significant decline in overall benchmark scores. For example, GPT‑6 Sol’s maximum effort score on the Artificial Analysis Intelligence Index remains high at 48, well above the median of 25, while Luna scores 37 against a median of 12. Cost savings are achieved despite a slight increase in output tokens per task, indicating efficiency gains are primarily from reduced operational expenses rather than model performance improvements.
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 for AI Deployment and Cost Efficiency
This development matters because lowering AI costs broadens the range of feasible applications, especially for organizations with tight budgets. The price cuts could accelerate adoption in areas like customer service, content generation, and research, where large-scale language models are integral. While the models’ benchmark scores remain stable overall, some minor regressions in knowledge tasks suggest users should test models within their specific workflows before full deployment. The move signals OpenAI’s focus on cost efficiency as a strategic priority, potentially reshaping how AI services are priced and consumed across industries.
As an affiliate, we earn on qualifying purchases.
OpenAI’s Cost-Reduction Strategy and Model Improvements
Two weeks prior to this announcement, OpenAI released GPT‑6 Astra, a high-end model emphasizing top-tier performance. The new Sol and Luna models are positioned as mid-tier options aimed at democratizing AI access by significantly reducing costs. The company states that these models benefit from advancements in caching and inference that lower operational expenses, passing savings directly to customers. Artificial Analysis’s evaluation confirms that these models maintain competitive benchmark scores, with some areas showing regression, likely due to tuning for cost and user experience rather than raw performance. Prior to this, OpenAI’s pricing for GPT‑5.6 models ranged from $4 to $20 per million tokens, making the new prices a notable shift towards affordability.
As an affiliate, we earn on qualifying purchases.
Unconfirmed Aspects of Long-Term Performance Impact
It is not yet clear how these models will perform over extended use cases or in real-world deployments, especially regarding the noted regressions in knowledge benchmarks. The long-term stability of the cost savings and whether future tuning might alter model performance or hallucination rates remain unknown. Additionally, the impact on competitive dynamics within the AI industry is still developing, as other providers may respond with their own pricing strategies.
enterprise AI deployment solutions
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps for Adoption and Model Evaluation
Organizations and developers are advised to test the models within their workflows to confirm suitability, especially for knowledge-intensive tasks. OpenAI is expected to release further details on caching tools and diagnostics, which will aid integration. Monitoring user feedback and independent evaluations over the coming months will clarify how these models perform in diverse applications and whether the cost reductions lead to broader AI deployment in industry sectors.
AI model performance monitoring tools
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 is priced at $2.00 per 1 million input tokens and $10.00 per 1 million output tokens. GPT‑6 Luna costs $0.10 for input and $0.50 for output per million tokens, representing a 50% reduction from previous prices.
Do the models’ benchmark scores indicate a performance drop?
Overall, benchmark scores remain stable or improved slightly, with some regressions in specific knowledge tasks. Independent analysis confirms the models maintain high performance levels, though some metrics show minor declines, likely due to tuning for cost efficiency.
What technical improvements enabled the price reductions?
OpenAI credits enhancements in caching and inference techniques, which reduce operational costs by enabling more effective reuse of context and decreasing the need for repeated computation, leading to lower prices.
Will these price cuts affect model quality or hallucination rates?
While hallucination rates have decreased, some quality regressions were observed in knowledge tasks. The models attempt fewer questions, which reduces errors but may also lower coverage in some cases. Users should evaluate models within their specific use cases.
What is the significance of these price changes for AI adoption?
Lower costs open opportunities for broader deployment in commercial and research settings, potentially accelerating AI integration into everyday workflows and reducing barriers for smaller organizations.
Source: ThorstenMeyerAI.com
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
