OpenAI GPT-6.1 Sol: Near-Astra Intelligence at One-Fifth the Price

What OpenAI Announced at DevDay 2026

At DevDay 2026 in San Francisco on September 29, OpenAI introduced GPT-6.1 Sol — an upgrade to GPT-6 Sol that, in OpenAI's words, "nearly matches GPT-6 Astra's intelligence on agentic coding, computer use, and professional work at one-fifth of Astra's standard input and output token prices." It is available now in ChatGPT Work, Codex and the API as gpt-6.1-sol, though OpenAI notes it is not yet available in regular ChatGPT chat.

The Pricing That Changes the Math

  • Standard input: $2 per million tokens — one-fifth of GPT-6 Astra's $10
  • Standard output: $10 per million tokens — one-fifth of Astra's $50
  • Cached input: $0.10 per million tokens — 95% less than standard input pricing and 50% less than GPT-6 Sol's cached input pricing
  • Long-context prompts (over 272K tokens) carry higher rates, per OpenAI's pricing

What the Benchmarks Say

Advertisement
Ad Placement (guide-detail-inline)

On DeepSWE v1.1, which evaluates complex software-engineering tasks in real codebases, OpenAI reports that GPT-6.1 Sol matches GPT-6 Astra at roughly one-fifth of the cost — while beating GPT-6 Sol's best score by 6.4 percentage points at lower reasoning effort and cost. On factual accuracy, OpenAI says the new model cuts the share of responses containing a factual error at low reasoning effort from 11.4% to 7.7% versus GPT-6 Sol, staying within 1.9% of GPT-6 Astra's error rate across all reasoning settings.

OpenAI also claims improvements in how the model handles constraints: it is more upfront about its limitations and more reliable at honoring user intent and safety restrictions — flagging broken tools, following explicit restrictions, and avoiding unauthorized outcomes in challenging evaluations. As with all vendor benchmarks, treat these as OpenAI's own measurements until independent evaluations land.

“We're introducing GPT-6.1 Sol, an upgrade to GPT-6 Sol that nearly matches GPT-6 Astra's intelligence on agentic coding, computer use, and professional work at one-fifth of Astra's standard input and output token prices.”

— OpenAI

Where You Can Use It (and Where You Cannot Yet)

SurfaceAvailabilityNotes
ChatGPT WorkAvailable nowPlus, Pro, Business, Enterprise and Edu plans
CodexAvailable nowIncludes the CLI and cloud workflows
API (gpt-6.1-sol)Available nowText + image in, text out
Regular ChatGPT chatNot yetOpenAI says availability is coming

The Bigger DevDay Picture

GPT-6.1 Sol landed alongside 20+ DevDay announcements, including Dots (always-on agents with their own cloud computer), the Ultrafast inference tier (up to 8x faster token generation in Codex at 6x standard pricing), the Agents API in public beta, and a $500/month Pro 500 plan. Notably, OpenAI did not ship the expected GPT-6.1 Astra upgrade — reporting indicates it was held back over safety concerns found in internal testing. The message of the event is clear: the frontier is now about cost-efficient, always-on agents, not just bigger models.

Frequently Asked Questions

How much cheaper is GPT-6.1 Sol than GPT-6 Astra?

Standard pricing is one-fifth of Astra's: $2 vs $10 per million input tokens and $10 vs $50 per million output tokens. Cached input is $0.10 per million, 95% below the standard input rate. For agentic workloads that reuse context, the real-world gap can be even wider.

Is GPT-6.1 Sol better than GPT-6 Sol?

OpenAI positions it as a direct upgrade: near-Astra intelligence on agentic coding, computer use and professional work, with a 6.4-point gain over GPT-6 Sol on DeepSWE v1.1, better factual accuracy, and half the cached-input price ($0.10 vs $0.20 per million tokens).

Can I use GPT-6.1 Sol in regular ChatGPT?

Not yet. OpenAI says the model is available in ChatGPT Work, Codex and the API (as gpt-6.1-sol), but it has not been rolled out to ordinary ChatGPT chat at launch.

What is the Ultrafast tier announced alongside it?

Ultrafast is a premium inference tier offering up to 8x faster token generation in Codex (up to 300 tokens per second) and up to 6x in the API, billed at 6x the standard price. It targets latency-sensitive agentic work.

Why was there no GPT-6.1 Astra?

OpenAI did not release a GPT-6.1 Astra upgrade at DevDay. Reporting attributed the decision to safety concerns raised during internal testing. GPT-6 Astra, launched September 3, remains OpenAI's top-tier model.

AdvertisementReserved ad placement · preview only