GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price
Points and comments are a snapshot, not live.
OpenAI releases GPT-6.1 Sol, nearly matching Astra's intelligence at one-fifth the cost.
GPT-6.1 Sol is an upgrade to GPT-6 Sol that approaches GPT-6 Astra's performance on coding, professional work, and computer use at one-fifth the price. Cached input costs $0.10 per million tokens. On DeepSWE v1.1, it matches Astra at roughly one-fifth the cost, eclipsing GPT-6 Sol's score by 6.4 points. On GDP.pdf, it scores higher than Opus 5.5 at less than half the cost per task. Factuality errors dropped from 11.4% to 7.7% at low reasoning effort. Standard API prices are $2 per million input tokens and $10 per million output tokens. Available in ChatGPT Work and Codex starting today.
What commenters are saying
Commenters note the naming confusion: a prior headline about safety concerns referred to GPT-6.1 Astra, not Sol. Some found GPT-6 Sol underwhelming compared to GPT-5.6 Sol, calling it worse in practice. Several see token price as the new battleground, with some arguing this signals approaching capability plateaus and investor panic, while others counter that progress hasn't plateaued and cite recent benchmark gains (e.g., Astra scoring ~24% on a math/code benchmark where earlier models scored 3-4%). A camp suggests the push for regulation is driven by IPO pressures and compute cost constraints, not capability limits.