OpenAI Unifies GPT-5.6 Sol Across All Paid Tiers and Drops Unlimited Free Chat

1. OpenAI Unifies GPT-5.6 Sol Across All Paid Tiers and Drops Unlimited Free Chat

On August 6, 2026, OpenAI announced that GPT-5.6 Sol now powers all chat modes for Plus and Pro subscribers, including both Instant and deep-reasoning responses, replacing the previous split where different modes ran on different underlying models. Free and Go users gain unlimited text chat on GPT-5.6 Luna starting August 7. OpenAI also added a reasoning-effort slider for Plus and Pro users, letting them dial how much compute ChatGPT applies per response. In high-stakes factuality evaluations covering finance, medicine, and law, GPT-5.6 Sol produced 68% fewer factually erroneous responses than GPT-5.5 Instant.

The strategic weight here is the access compression. OpenAI has historically used model tiers to segment paid from free users and to differentiate Plus from Pro. Collapsing Instant and deep-reasoning onto a single model removes one of those internal friction points, making the paid experience more coherent. More aggressively, putting GPT-5.6 Luna on the free tier with unlimited text chat raises the floor for what a zero-cost AI assistant looks like. That directly undercuts Anthropic's Claude, which keeps its strongest free-tier model behind usage caps, and pressures Google's Gemini free tier to match capability rather than just volume. The 68% factuality improvement figure, specifically called out for finance, medicine, and law, signals OpenAI is targeting professional use cases where accuracy failures are costly.

The reasoning-effort slider is worth watching as a product pattern. Giving users direct control over compute per query is a pricing and UX experiment simultaneously. If it sticks, expect Anthropic and Google to follow with similar controls. The broader arc: OpenAI is treating model access as a distribution weapon, not just a revenue ladder.

Source: @OpenAI on X