NEWS · MODELS · #814
OpenAI's GPT-6 Sol and GPT-6 Luna now generally available on Amazon Bedrock
OpenAI's GPT-6 Sol and GPT-6 Luna are now generally available on Amazon Bedrock, running on Bedrock's high-performance inference engine and offered at lower API pricing than their GPT-5.6 predecessors. GPT-6 Sol is positioned for demanding, recurring development and operations tasks with improved coding, reasoning and factuality (OpenAI reports ~50% fewer factual mistakes vs GPT-5.6 Sol), while GPT-6 Luna is optimized for high-volume, repeatable workloads like extraction, classification and summarization; Bedrock supports explicit prompt caching plus IAM-based access control and invocation auditing.
KEY POINTS
- OpenAI's GPT-6 Sol and GPT-6 Luna are now generally available on Amazon Bedrock, running on Bedrock's high-performance inference engine and offered at lower API pricing than their GPT-5.6 predecessors.
- GPT-6 Sol is positioned for demanding, recurring development and operations tasks with improved coding, reasoning and factuality (OpenAI reports ~50% fewer factual mistakes vs GPT-5.6 Sol), while GPT-6 Luna is optimized for high-volume, repeatable workloads like extraction, classification and summarization; Bedrock supports explicit prompt caching plus IAM-based access control and invocation auditing.
- This widens practical production use of GPT-6 by pairing improved model quality and lower cost with Bedrock's performance, caching, and governance features, enabling broader high-volume and recurring workflows.
DECISION BRIEF
CONFIDENCE
WHAT CHANGED
OpenAI released two GPT-6 tier models (GPT-6 Sol and GPT-6 Luna) and they are now generally available on Amazon Bedrock; the release pairs model updates with lower API pricing and runtime/caching improvements.
WHY NOW
The combination of GPT-6 model updates, roughly halved API token pricing, and availability on Bedrock’s high-performance inference platform increases the short-term practicality of using GPT-6 for recurring and high-volume production tasks.
WHO IS AFFECTED
- Organizations deploying LLMs on Amazon Bedrock - Developers and SREs using models for recurring development/operations tasks (code implementation, debugging, refactoring, reviews) - Teams running high-volume, repeatable workloads (extraction, classification, summarization) - Users of ChatGPT Work / Codex and API customers (availability differs by tier)
CONFIRMED
- GPT-6 Sol and GPT-6 Luna are generally available on Amazon Bedrock. - Bedrock offers an inference engine described as built for high performance, security, and reliability at scale. - OpenAI and AWS say the new GPT-6 Sol and Luna are offered at significantly lower API pricing than their GPT-5.6 predecessors (described as a ~50% price cut). - OpenAI reports that, on an internal factuality evaluation, GPT-6 Sol makes about half as many factual mistakes as GPT-5.6 Sol. - OpenAI positions GPT-6 Sol for demanding recurring development and operations tasks (coding, debugging, refactor, review) and GPT-6 Luna for high-volume, repeatable workloads (summarization, extraction, short answers). - Reported example token prices from one media excerpt: GPT-6 Sol $2 per million input tokens and $10 per million output tokens; GPT-6 Luna $0.10 per million input tokens and $0.50 per million output tokens. - OpenAI reports improvements in caching (including a reported 90% discount on cached input tokens and new prompt-caching tools) and attributes part of the price drop to caching/inference improvements. - Availability notes in reporting: models are available in ChatGPT Work and Codex for paid subscribers; Luna has wider desktop/free-user availability in some channels; API names include gpt-6-sol and gpt-6-luna; ChatGPT rollout is gradual.
UNCERTAIN
- Independent, third‑party benchmark consensus on absolute performance gains vs GPT-5.6 and vs competitors is mixed or incomplete in the supplied excerpts: some reporting emphasizes substantial cost-per-task advantages and improved benchmark scores, while another outlet characterizes the changes as mostly a price cut with little performance gain. - Long‑term real‑world production reliability, measured outside OpenAI’s internal evaluations, is not documented in the supplied excerpts. - Exact Bedrock-side governance details cited in the event summary (explicit prompt caching implementation, IAM-based access control, invocation auditing) are not fully documented in the provided source excerpts or are not quoted verbatim across all supplied snippets — confirmation of those specific Bedrock features in these excerpts is limited. (source 959; note: some governance/caching claims come from OpenAI reporting rather than independent Bedrock documentation in the excerpts) - Full benchmark numbers and methodology (e.g., FrontierCode, AutomationBench, OSWorld 2.0) are only partially excerpted; the supplied text truncates some results, so comprehensive comparison data are missing.
WHAT TO WATCH
Observable signals to monitor next (concrete): - AWS/Bedrock documentation or blog updates listing specific governance features (IAM controls, invocation auditing, Bedrock-side prompt caching details). (watch for updates from AWS) - Independent benchmark releases (AutomationBench, FrontierCode, OSWorld 2.0) or third‑party evaluations comparing GPT-6 Sol/Luna to GPT-5.6 and competitor models. - Usage/availability announcements from OpenAI or AWS showing uptake on Bedrock, enterprise case studies, or new channel rollouts in ChatGPT/ChatGPT Work/Codex. - Official pricing pages or API docs confirming token prices, cached-token discounts, and the gpt-6-sol / gpt-6-luna API identifiers. - Documentation or dashboards for prompt-caching tools and diagnostics (to verify the claimed 90% cached-input discount and cache behavior changes).
WHY IT MATTERS
This widens practical production use of GPT-6 by pairing improved model quality and lower cost with Bedrock's performance, caching, and governance features, enabling broader high-volume and recurring workflows.
EVIDENCE MAP
4Editorial claims linked to specific sources, with support, contradiction and context shown separately.
GPT-6 Sol and GPT-6 Luna are generally available on Amazon Bedrock.
SUPPORTEDChecked 2026-09-24 · 1 supporting
Both GPT-6 Sol and GPT-6 Luna are offered at lower API pricing than their GPT-5.6 predecessors (about a 50% price cut reported).
SUPPORTEDChecked 2026-09-24 · 3 supporting
TechCrunch reports a price drop to half the cost of the 5.6 series, attributing it to caching and inference improvements.
SUPPORTS AWS Machine LearningMEDIA · 2026-09-24AWS blog says both models come at significantly lower API pricing than GPT-5.6 predecessors.
SUPPORTS The DecoderMEDIA · 2026-09-24The Decoder states a 50% price cut and gives example token prices for Sol and Luna.
OpenAI reports that GPT-6 Sol makes roughly half as many factual mistakes as GPT-5.6 Sol on an internal factuality evaluation.
SUPPORTEDChecked 2026-09-24 · 2 supporting
TechCrunch repeats OpenAI's claim that GPT-6 Sol makes about half as many mistakes as its predecessor on an internal evaluation.
SUPPORTS AWS Machine LearningMEDIA · 2026-09-24AWS blog quotes OpenAI's internal factuality evaluation finding GPT-6 Sol made approximately half as many factual mistakes as GPT-5.6 Sol.
On Amazon Bedrock, the models run on an inference engine described as built for high performance, security, and reliability at scale.
SUPPORTEDChecked 2026-09-24 · 1 supporting
AWS blog states the models run on an inference engine built for high performance, security, and reliability at scale.
CONTEXT OpenAIMEDIA · 2026-09-24OpenAI's announcement introduces the models but the excerpt does not detail Bedrock's inference engine.
SOURCES & TIMELINE
4Meet GPT-6 Sol and Luna, two models that bring frontier intelligence to everyday work with different balances of capability and cost.
Earlier this month, OpenAI launched GPT-6 Astra , which it heralded as its most powerful and capable model yet, and the “world’s best model” for a variety of activities, including computer work and coding. Now, the lab is expanding the GPT-6 generation with updated versions of the smaller Sol and Luna models. “GPT-6 Astra introduced a new generation of intelligence; these models extend its benefits by making that in…
GPT-6 Sol and GPT-6 Luna are now generally available on Amazon Bedrock, giving you more options to match intelligence and efficiency to each workload. The value of AI at scale depends on two dimensions: what a model can do and how often you can put it to use. Greater intelligence expands the complexity a model can handle, from subtle coding problems to multistep processes across tools. Efficiency determines how broa…
With GPT-6 Sol and Luna, OpenAI adds two cheaper models to its lineup that match their predecessors' performance at half the token price and aim to rival some of Anthropic's more expensive models. The biggest change is a 50 percent price cut compared with GPT-5.6 Sol and Luna. GPT-6 Sol now costs $2 per million input tokens and $10 per million output tokens, while Luna comes in at $0.10 for input and $0.50 for outpu…