> ## Content Index > Fetch the complete content index at: https://eazzytechnews.ghost.io/llms.txt > Use this file to discover other available public pages before exploring further. # OpenAI gives developers a cheaper GPT-6 ladder - URL: https://eazzytechnews.ghost.io/openai-gpt6-sol-luna-api/ - Published: 2026-09-23T23:55:03.000Z - Updated: 2026-09-25T15:34:28.000Z - Description: OpenAI released GPT-6 Sol and GPT-6 Luna in the API, creating a cheaper GPT-6 ladder for teams that do not need Astra on every request. - Author: Collins Anfo - Tags: OpenAI, GPT-6, Developer Tools, AI Pricing, Cloud AI OpenAI has added GPT-6 Sol and GPT-6 Luna to its API, giving developers a lower-cost GPT-6 path for jobs that do not need the full Astra tier. The [OpenAI API changelog lists the release on September 22](https://developers.openai.com/api/docs/changelog?ref=eazzytechnews.ghost.io). Both models accept text and image inputs and generate text through the Responses API and Chat Completions. The change turns GPT-6 into a routing stack: Astra for the hardest end-to-end work, Sol for regular complex reasoning, and Luna for high-volume tasks where cost matters most. ![OpenAI Developers changelog graphic for the API docs changelog](https://storage.ghost.io/c/09/53/09539035-af35-486d-b626-5b7c2dda5b1d/content/images/2026/09/openai-changelog-og.png) OpenAI's developer documentation image identifies the API docs changelog that lists GPT-6 Sol and GPT-6 Luna. Source: OpenAI Developers. The pricing difference is large enough to change how teams design model routing. For prompts up to 272,000 input tokens, OpenAI lists GPT-6 Sol at $2 per million input tokens, $0.20 per million cached input tokens and $10 per million output tokens. GPT-6 Luna is listed at $0.10 per million input tokens, $0.01 per million cached input tokens and $0.50 per million output tokens. Longer prompts, cache writes and other processing tiers still require the pricing page, so the changelog should be read as the standard starting point rather than a complete invoice model. ## The release makes routing more important The practical question is not whether Sol or Luna is "better" in the abstract. It is which work should be routed to which model. A support classifier, short extraction task or high-volume assistant response may not need Astra. A coding agent that has to understand an existing system, call tools and recover from failed tests may need more reasoning headroom. A document workflow might move between models as it drafts, checks and rewrites. That gives engineering teams a concrete migration checklist: - Separate high-risk reasoning tasks from routine text and image-input tasks. - Log fresh input, cached input, output tokens and total spend by route. - Compare success rates before moving a workflow from Astra or an older model to Sol or Luna. - Use cached input pricing only where the application actually reuses context. The cache line is easy to underestimate. A long-running assistant or agent often rereads the same instruction set, project context and intermediate results. If cache reuse works, a lower cached-input price can be as important as the headline input rate. If the application constantly changes context and misses the cache, the savings will be smaller. ## AWS makes it an enterprise distribution story AWS moved quickly behind the release. In a same-day post, [Amazon said GPT-6 Sol and GPT-6 Luna are generally available on Amazon Bedrock](https://aws.amazon.com/blogs/machine-learning/bring-more-intelligence-to-everyday-work-with-gpt-6-sol-and-gpt-6-luna-on-amazon-bedrock/?ref=eazzytechnews.ghost.io). AWS frames Sol for recurring complex tasks such as software development, operations and analytics, while Luna is positioned for focused, economical tasks at higher volume. ![AWS Artificial Intelligence Blog graphic for GPT-6 Sol and GPT-6 Luna on Amazon Bedrock](https://storage.ghost.io/c/09/53/09539035-af35-486d-b626-5b7c2dda5b1d/content/images/2026/09/aws-bedrock-gpt6-sol-luna-featured.png) AWS published this article graphic for its Bedrock announcement covering GPT-6 Sol and GPT-6 Luna availability. Source: AWS. That cloud availability changes procurement more than model capability. Companies already standardizing model access through Bedrock can test OpenAI's new tiers without building a separate vendor path for every internal team. It also places Sol and Luna inside the broader enterprise controls that Bedrock customers already use for identity, billing and model access. The buyer risk is familiar: cheaper models can produce expensive mistakes if they are routed into work they cannot reliably complete. A low token bill is not useful if the system needs repeated retries, human cleanup or escalation after bad tool calls. The right test is therefore task cost, not token price alone. Teams should measure completed work per dollar, with quality gates, retries and review time included. For OpenAI, Sol and Luna fill an obvious gap below Astra. For developers, the news is more operational: the GPT-6 stack now has enough price spread that model routing becomes part of product architecture from the start. **Author:** [Collins Anfo, a digital product builder. My write-ups cover technology and artificial intelligence, from software, intelligent agents and robotics to chips, computing infrastructure, cybersecurity, consumer devices and space technology. I also cover technology-driven discoveries across mathematics, science and other research fields, product releases and announcements, and the business, finance, investment, governance and real-world adoption of technology.](https://collins-anfo-portfolio-2026.collinsanfo24.chatgpt.site/?ref=eazzytechnews.ghost.io) **AI assistance disclosure:** AI tools assisted with the research and drafting of this article. Material claims are linked to their sources for independent verification.