GPT-Live is here — and GPT-5.6 is coming.
OpenAI shipped two announcements in one week. One you can use today: GPT-Live, a new generation of voice models rolling out to everyone. One you can't yet: GPT-5.6, a three-model family in limited preview. Here's what each actually is, what's real today, and what to ignore until it ships.
01 GPT-Live: voice that finally converses
Announced July 8, GPT-Live is built on a full-duplex architecture — it listens and speaks at the same time, like a person on a phone call. In practice that means it can react while you're mid-sentence ("mhmm," "yeah"), handle quick back-and-forth without the walkie-talkie pauses, and — the underrated part — stay quiet when you need a moment to think.
It's also the smartest voice model OpenAI has shipped, with a trick: for questions needing web search or deeper reasoning, GPT-Live delegates to a frontier model behind the scenes (GPT-5.5 at launch) and folds the answer back into the conversation.
| Version | Who gets it | Status |
|---|---|---|
| GPT-Live-1 | Default voice model for paid tiers (Go, Plus, Pro) | Rolling out globally now |
| GPT-Live-1 mini | Free tier | Rolling out globally now |
Try it today: open voice mode in the ChatGPT app (iOS, Android, or chatgpt.com) and interrupt it mid-answer — then ask a question that needs current web info and notice the "let me check" delegation. If your voice mode still feels walkie-talkie, the rollout hasn't reached you yet; it's arriving over days, not months. Our voice mode lesson covers the workflows once you have it.
02 GPT-5.6: Sol, Terra, and Luna — decoded
GPT-5.6 introduces a naming scheme worth learning now, because you'll be choosing between these soon:
| Model | Built for | Think of it as |
|---|---|---|
| Sol | Deep reasoning — programming, biological analysis, cybersecurity | The heavyweight |
| Terra | Best balance of performance and cost | The daily driver |
| Luna | Speed | The sprinter |
03 The quieter ships worth knowing
Two smaller July updates that change real workflows:
Memory doubled. Plus and Pro users now get roughly 2x memory capacity — ChatGPT can hold more durable context about you and your work. If you set up memory ages ago and hit the cap, it's worth a pruning pass now that there's room. (How to do that well: memory & custom instructions.)
Lockdown Mode. An opt-in security setting now available to all logged-in users. It restricts network-enabled capabilities — live browsing, deep research, agent mode, file downloads — for people whose threat model demands it (journalists, executives, anyone handling sensitive material). Most users shouldn't turn it on; the right users really should.
04 What to actually do with this
What you can do now
- Test GPT-Live in voice mode — interrupt it, let it pause, push a web question at it
- Learn the Sol / Terra / Luna split now; choose nothing until it's actually in ChatGPT
- Revisit your saved memories — the cap just doubled for Plus/Pro
- Know Lockdown Mode exists; enable it only if your threat model calls for it
- Ignore anyone claiming GPT-5.6 access inside ChatGPT this week
This week's challenge
Run one full commute (or a 15-minute walk) as a GPT-Live working session: brief it on a real problem out loud, let it push back, and ask for a written summary at the end. Compare that summary against what a typed session gives you. Voice stopped being a demo this week — find out if it earns a slot in your routine.