The Information Machine
Following·Day 2·first covered 29 Sep 2026·5 sources

OpenAI Launches GPT-6.1 Sol at 1/5th Astra Price, Safety Data Published

The gist

The pricing structure, at one-fifth the cost of GPT-6 Astra with deep cache discounts, positions GPT-6.1 Sol as an accessible option for high-volume or latency-sensitive workloads. The safety card data shows a substantially higher exploit success rate than prior Sol models, which OpenAI acknowledges while noting it remains below GPT-6 Astra.

The full picture

OpenAI released GPT-6.1 Sol on September 29, 2026 at DevDay, priced at one-fifth of GPT-6 Astra with a 95% cache read discount on cached reads. The model slots below GPT-6 Astra and above GPT-6 Luna in the GPT-6 family, released one week after GPT-6 Sol and 26 days after GPT-6 Astra. Sam Altman described it as 'an absolute workhorse' delivering 'near Astra intelligence.' Standard input and output token rates remain unchanged from GPT-6 Sol, but cached reads cost less and the tool-calling contract changed. Tool calling is supported only through the Responses API, and ChatGPT Chat access was not yet available at announcement. OpenAI also launched an Ultrafast service tier delivering up to approximately 300 tokens per second across GPT-6-class models at 6x the corresponding model's standard rate, available immediately for GPT-6 Astra and expected for GPT-6.1 Sol within days of announcement. Safety evaluations show GPT-6.1 Sol outperforms earlier Sol models on most health and safety benchmarks while trailing GPT-6 Astra on exploit capability.

How it developed
29 September 2026

OpenAI launches Ultrafast service tier at ~300 tokens/sec for GPT-6-class models at 6x standard rate

Sources
The daily email

Want this in your inbox?

I send one email each morning with the stories that moved. If you would rather just read here, that works too.

Subscribe free