GPT-6 Sol and Luna
OpenAI's cost-optimized GPT-6 models (Sol for difficult work tasks, Luna for cheaper high-volume use), launched in September 2026 with fewer mistakes at roughly half the price of the GPT-5.6 tier.
Last verified · 2026-10-23 · by Moe Ameen
Published 2026-09-22 · Updated
Short answer
GPT-6 Sol and GPT-6 Luna are OpenAI's cheaper GPT-6 models, launched in September 2026 after GPT-6 Astra. Sol costs $2 per million input tokens and $10 output; Luna costs $0.10 and $0.50, half of GPT-5.6 promotional prices. OpenAI says Sol makes about half as many factual mistakes as its predecessor. Both are text and reasoning models that do not publish anything.
Key takeaways
- Sol: $2 in / $10 out per million tokens. Luna: $0.10 in / $0.50 out.
- Both are 50% cheaper than GPT-5.6 promotional pricing, per OpenAI.
- OpenAI says Sol makes about half as many factual mistakes as GPT-5.6 Sol.
- They are in ChatGPT Work and Codex for paid plans; Free and Go users get Luna in the desktop app.
- Use Luna for bulk copy and Sol for harder drafts, then Kompozy to publish.
What are GPT-6 Sol and Luna?
GPT-6 Sol and GPT-6 Luna are the cost-optimized models in OpenAI's GPT-6 family, launched in September 2026 after the flagship GPT-6 Astra earlier the same month. OpenAI positioned them around a plain promise rather than a new capability: lower cost and fewer mistakes.
Both ship at roughly half the price of their GPT-5.6 namesakes — OpenAI listed Sol at $2 per million input tokens and $10 per million output tokens (down from $4 and $20), and Luna at $0.10 and $0.50 (down from $0.20 and $1.20) — and credited improvements in caching and inference. OpenAI measured the 50% cut against GPT-5.6's promotional pricing.
The two models are built for different jobs. OpenAI says Sol can take on difficult work tasks with higher usage limits and lower cost. Luna is the cheaper tier, and OpenAI says that at higher effort levels it matches GPT-5.6 Sol on factuality at about a hundredth of the cost. On OpenAI's internal factuality evaluation, which it describes as drawn from de-identified real conversations where users flagged model mistakes, OpenAI said Sol makes roughly half as many mistakes as GPT-5.6 Sol — approaching flagship Astra reliability at much lower cost — and said Sol matches Claude Fable 5.1 xhigh on the FrontierCode coding benchmark at much lower cost.
Both are available through the OpenAI API and are rolling out in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise, and Edu users, with Luna also reaching the desktop app and free tiers. One thing to be clear about: Sol and Luna are text and reasoning models. They write, reason, and code — they generate no images, no video, no captions, and they publish to no platform. Treat exact pricing, benchmark figures, and rollout details as a launch-window snapshot and confirm them against OpenAI's own pages.
What can you make with GPT-6 Sol and Luna?
- Reasoned first drafts and outlines from a rough brief — a video script, a blog structure, a thread, a newsletter body (Sol)
- High-volume bulk copy — dozens of caption variations, per-platform rewrites, and hook options generated cheaply (Luna)
- Summaries and information extraction over documents, transcripts, or notes you provide (Luna)
- Code, debugging help, and technical explainers for developer-leaning audiences (Sol)
- Research synthesis that turns a messy source into a clean set of post angles
How does Kompozy turn GPT-6 Sol and Luna drafts into content?
Sol and Luna are built for two different jobs, and that split maps cleanly onto how Kompozy uses a model. Luna is the cheap, high-volume workhorse — ideal for generating the bulk copy a content week needs: dozens of caption variations, per-platform rewrites, a source summarized into ten post angles. Sol is the one you reach for when the draft has to reason — a tight Persona Short script, a structured blog outline, a synthesis of a rambling transcript. Neither, though, turns any of that into something you can post. That is the handoff Kompozy is built for.
Concretely: draft in Sol or Luna, then bring the text into Kompozy as your source. Kompozy rewrites it through a Persona Brief so it reads like you, then generates the finished formats a model can't — a captioned Persona Short or a HeyGen avatar video narrating the script, a brand-exact Carousel built from the outline, Quote Graphics pulling the sharpest lines, Clipped Shorts from a longer cut, per-platform Text Posts, a full Blog Article, and an Email Newsletter across Kompozy's output formats. Then Autopilot schedules and publishes the set across the eight social platforms plus blog and email behind a per-post review. Because Luna is so cheap you can afford to over-generate raw copy and let Kompozy shape the best of it; because Sol is more accurate, the scripts it hands off need less fixing. The model thinks and drafts; Kompozy makes and ships.
- Pick the right model for the draft — Luna for high-volume, tightly-scoped copy (captions, summaries, angle lists), Sol for the reasoning-heavy piece (a script, an outline, a synthesis).
- Bring the draft into Kompozy as a source and set a Persona Brief so voice, look, and banned words carry across every output.
- Generate the media the model can't: a captioned Persona Short or avatar video, brand-exact carousels, quote graphics, clipped shorts, per-platform text posts, a blog, and a newsletter.
- Review each piece in the per-post queue, letting Kompozy reframe it for TikTok, Reels, Shorts, X, LinkedIn, and the rest.
- Schedule and publish the full set across the eight social platforms plus blog and email on Autopilot — one cheap, accurate draft turned into a published content week.
What does this look like in Kompozy?
An actual Kompozy output in the persona tweet format. It was made for a different subject than GPT-6 Sol or Luna, and it shows the finished format you would get.
Who is this for, and who should skip it?
Best for:
- Teams that want near-flagship quality at lower token cost (Sol)
- High-volume summarizing, extraction, and caption variants (Luna)
- Developers already building on the OpenAI API or Codex
Not ideal for:
- Work that needs the very best results, where OpenAI still points to Astra
- Anyone who needs image, video, or audio output
- Creators who want posts published without a second tool
Frequently asked questions
What are GPT-6 Sol and Luna?
They are the cost-optimized models in OpenAI's GPT-6 family, launched in September 2026 after the flagship GPT-6 Astra. Sol is the stronger of the two for difficult work tasks; Luna is the cheaper tier for running at volume. OpenAI positions both around lower cost and fewer mistakes.
What is the difference between GPT-6 Sol and Luna?
Sol is the higher-capability model, and OpenAI pitches it for difficult work tasks with higher usage limits. Luna is the cheaper tier at $0.10 per million input tokens versus Sol's $2. At higher effort levels, OpenAI says Luna matches GPT-5.6 Sol on its factuality evaluation at about a hundredth of the cost.
Can GPT-6 Sol or Luna generate video or images?
No. Both are text and reasoning models — they draft, code, and analyze but produce no video, images, or audio, and publish to no platform. To turn their drafts into avatar video, carousels, quote graphics, and other finished media, pair them with a content engine like Kompozy that generates those formats and publishes them.
How do I turn GPT-6 Sol or Luna output into social posts?
Use Luna to generate bulk copy or Sol to draft the reasoning-heavy piece, then bring that into Kompozy. Kompozy rewrites it in your voice via a Persona Brief, generates Shorts, carousels, quote graphics, text posts, a blog, and a newsletter, and schedules and publishes them across the eight social platforms plus blog and email.
How much do GPT-6 Sol and Luna cost?
OpenAI listed Sol at $2 per million input tokens and $10 per million output tokens, and Luna at $0.10 and $0.50 , about half the promotional prices of their GPT-5.6 counterparts. Confirm current figures on OpenAI's own pages, as pricing and availability vary by plan and region.
How do we know this?
We checked every figure and product claim on this page against the sources listed below on 2026-10-23, and removed or corrected anything they did not support. OpenAI's post carries no exact publish day, so the page now says September 2026. Kompozy plan and platform details come from kompozy.io/pricing and kompozy.io.
Sources
- OpenAI: Introducing GPT-6 Sol and Luna (read 2026-10-23)
- Kompozy: Home (read 2026-10-23)
Related tools
- GPT-6 Astra — OpenAI's flagship reasoning and computer-use model — strong at drafting, coding, science, and operating software, and the successor to GPT-5.6 Sol.
- GPT-5.6 Sol — OpenAI's flagship GPT-5.6 tier — a frontier reasoning, writing, and tool-orchestration model that reads reference images and drives multi-tool creative pipelines, but generates no media itself.
- ChatGPT — OpenAI's AI assistant for writing, ideas, and images — which, since late July 2026, refuses direct requests to write "in the style of" a specific named author.
- Claude Opus 5 — Anthropic's frontier Claude model — deep reasoning, agentic coding, vision, and a 1M-token context, positioned near Fable 5 intelligence at half the cost. A text-output model, not an image, audio, or video generator.
- ChatGPT Pro — OpenAI's top individual ChatGPT plan — $200/month for the highest message limits and priority access to its most advanced models and tools.
- Grok 4.7 — xAI's most powerful model for coding and knowledge work — a larger base than Grok 4.6, tuned for harder multi-hour tasks and careful self-verification.