OpenAI removed text message limits for free ChatGPT users starting the week of August 10, 2026, marking the biggest change to its free tier since launch.

Free accounts can now send unlimited text messages to ChatGPT but remain restricted to GPT-5.6 Luna, the smallest model in OpenAI's GPT-5.6 family. Access to larger models like GPT-5.6 Terra and Sol requires a paid subscription.

The free tier includes basic reasoning mode through the Think button, though without the slider controls available to subscribers. Users get web search, image generation, and voice mode using the Live-1 mini model.

What stays limited

File uploads remain capped at 500MB through ChatGPT's Library feature. Weekly limits still apply to image generation, file processing, Deep Research, voice mode, and OpenAI's Codex and Work tools.

The company stopped sharing exact usage numbers publicly in July 2026, moving from five-hour rolling windows to weekly limits. Users can check remaining quotas in Settings under the Usage section.

Free accounts see sponsored ads below ChatGPT responses, often recommending products related to user prompts. OpenAI says ads won't appear for sensitive topics like health or mental wellbeing.

Training opt-out available

OpenAI automatically enrolls free users in model training but allows opt-out through Data Controls settings. The $8 monthly Go tier still shows ads, while the $20 Pro plan removes them entirely.

Context windows and memory features remain more limited on free accounts compared to paid tiers. The change puts OpenAI's free offering closer to Anthropic's Claude approach, though specific limit comparisons aren't publicly available.

The move comes as competition intensifies among AI labs for user adoption, with unlimited text access removing a key friction point for ChatGPT's free tier.