OpenAI's upgrade to GPT-6 Sol — near-flagship intelligence for agentic coding, computer use, and professional work at one-fifth the price of GPT-6 Astra, launched September 29, 2026.
Last verified · 2026-09-29 · by Moe Ameen
GPT-6.1 Sol is OpenAI's upgrade to GPT-6 Sol, announced September 29, 2026 at DevDay — a week after GPT-6 Sol itself launched. The claim OpenAI leads with is value, not a new headline capability: it says GPT-6.1 Sol reaches nearly the same level of intelligence as its flagship GPT-6 Astra on agentic coding, computer use, and professional work, while costing about one-fifth as much for standard tokens.
On price it holds GPT-6 Sol's rates — $2 per million input tokens and $10 per million output tokens — and halves cached input to $0.10 per million. Because GPT-6 Astra sits at $10 input and $50 output, GPT-6.1 Sol's standard prices are exactly one-fifth of the flagship's. On OpenAI's internal factuality measure the model's error rate lands within roughly 1.9% of Astra across reasoning settings, and at low reasoning effort OpenAI reported it dropping from 11.4% on GPT-6 Sol to 7.7%. It also cited parity with Astra on a software-engineering benchmark and a near-match on a computer-use test at maximum reasoning.
It is aimed at builders first. Access is through the API as `gpt-6.1-sol`, and it is rolling out inside ChatGPT Work and Codex for Plus, Pro, Business, Enterprise, and Edu users — not yet in the standard Chat surface. OpenAI said an "Ultrafast" variant that generates tokens much faster, at a higher price, is coming in the days after launch.
One thing to be clear about: GPT-6.1 Sol is a text and reasoning model. It writes, reasons, codes, executes multi-step workflows, and reads documents — but it generates no images, no video, no captions, and it publishes to no platform. Treat exact pricing, benchmark figures, and rollout details as a launch-window snapshot and confirm them against OpenAI's own pages.
Read GPT-6.1 Sol for what it is best at — accurate, agentic, professional-grade reasoning — and its value to a creator becomes obvious: it is the sharpest, cheapest *drafting and research brain* you can put at the front of a content workflow. It shines at turning a rambling transcript into clean post angles, chaining a topic into a script and then a caption set, and reading a report to pull the three ideas worth posting. Because its factual-error rate now approaches the flagship's, the drafts it hands off need less fixing; because it is priced at one-fifth of Astra, you can afford to over-generate and keep only the best. What it cannot do is make any of that into something you can post — it returns text and stops. That handoff is the whole reason [Kompozy](/) exists.
Run it as a clean two-stage line. Draft and synthesize in GPT-6.1 Sol (in Codex or ChatGPT Work), then drop the script, outline, or angle list into Kompozy's Quick Ingest as the seed for a content unit. Kompozy rewrites it through a [Persona Brief](/glossary/persona-brief) so it reads like you, then generates the finished formats a model can't: a captioned [Persona Short](/glossary/persona-shorts) or a HeyGen avatar video narrating the script, [Persona Frames](/glossary/persona-frames) composited into a brand-exact template, a brand-exact [Carousel](/glossary/hyperframes) built from the outline, Quote Graphics pulling the sharpest lines, a [Clipped Short](/glossary/clipped-short) from a longer cut, per-platform Text Posts, a full Blog Article, and an Email Newsletter — across [18 formats](/glossary/output-buckets). Then [Autopilot](/glossary/autopilot) schedules and publishes the set across the eight social platforms plus blog and email behind a per-post review. The model does the thinking and drafting; Kompozy does the making, branding, and shipping it was never built to do. For related models see [GPT-6 Sol and Luna](/ai-tools/gpt-6-sol-and-luna) and the flagship [GPT-6 Astra](/ai-tools/gpt-6-astra).
GPT-6.1 Sol is OpenAI's upgrade to GPT-6 Sol, announced September 29, 2026 at DevDay. OpenAI says it nearly matches the flagship GPT-6 Astra on agentic coding, computer use, and professional work while costing about one-fifth as much for standard tokens. It is a text and reasoning model, available in the API and in ChatGPT Work and Codex.
OpenAI listed it at $2 per million input tokens and $10 per million output tokens — the same standard rate as GPT-6 Sol — with cached input halved to $0.10 per million. Against GPT-6 Astra's $10 input and $50 output, those standard prices are exactly one-fifth. Confirm current figures on OpenAI's own pricing pages, as they vary by plan and region.
No. It is a text and reasoning model — it drafts, codes, reads documents, and executes multi-step workflows, but produces no video, images, or audio and publishes to no platform. To turn its drafts into avatar video, carousels, quote graphics, and other finished media, pair it with a content engine like Kompozy that generates those formats and publishes them.
Draft the reasoning-heavy piece in GPT-6.1 Sol, then bring it into Kompozy. Kompozy rewrites it in your voice via a Persona Brief, generates Persona Shorts, carousels, quote graphics, text posts, a blog, and a newsletter from the one source, and schedules and publishes them across the eight social platforms plus blog and email.