Models and effort
Which model answers, how hard it thinks, and what that costs you.
Models
Every model below is included in every paid plan. The Free plan has the lighter ones: Auto, Claude Sonnet 5.5 and Sonnet 5, Claude Haiku 4.5, GPT-6 Luna and GPT-5.6 Luna, GLM-5.2 and PocketPhantom Noir; the others show a lock there and lead to the plans. There is nothing to set up and no API key of your own to bring. Pick one in the model pill at pocketphantom.com — that picker is the whole list today, because the web app is the whole product for now. Auto lets us pick; today that is GLM-5.3, or Claude Opus 5.5 for a message with files, though one with only pictures may go to PocketPhantom Noir instead. On the Free plan Auto stays with the lighter models: GLM-5.2, or Claude Sonnet 5.5 for a message with files.
Anthropic
| Model | Good for | Speed |
|---|---|---|
| Claude Opus 5.5 | Serious work: long reasoning, tricky code, careful writing. Released September 2026. | Slow |
| Claude Opus 5 | The previous Opus. | Slow |
| Claude Fable 5.2 | Anthropic's most capable model, for the hardest problems and long jobs. | Slowest |
| Claude Fable 5.1 | The previous Fable. | Slowest |
| Claude Sonnet 5.5 | Most things. Sharp and fast. | Fast |
| Claude Sonnet 5 | The previous Sonnet. | Fast |
| Claude Haiku 4.5 | Quick questions, short edits, lots of small asks. | Quickest |
OpenAI
| Model | Good for | Speed |
|---|---|---|
| GPT-6 Astra | OpenAI's strongest: computer use, browsing, software engineering, long documents (a million tokens of context). Released September 2026, rolling out in phases. | Slow |
| GPT-6.1 Sol | OpenAI's newest, from late September 2026: reasoning, code, research with the web, long documents (a million tokens of context), at a fifth of Astra's price. | Slow |
| GPT-6 Luna | The small, cheap GPT-6: focused everyday jobs, done quickly. | Fast |
| GPT-5.6 Luna | One of three 5.6 models. OpenAI has said nothing about what sets them apart, so we claim nothing either: reasoning, code, long documents. | Slow |
| GPT-5.6 Sol | The second of the three, same family and same generation. | Slow |
| GPT-5.6 Terra | The third. They are peers; take whichever answers you best. | Slow |
| GPT-5.5 | The previous flagship; still strong everywhere. | Slow |
| GPT-5.5 Pro | The 5.5 that thinks longest before it answers, and waits longest for it. | Slowest |
GLM, uncensored
GLM is an open model family from Zhipu AI; ours are builds with the refusal training taken out. They answer the questions the big labs decline — security research, fiction that goes somewhere dark, blunt medical or legal questions, a subject somebody decided was awkward. They sit in the picker with the rest, marked with a blue Uncensored tag. GLM-5.2 and PocketPhantom Noir are in every plan, the Free plan too; GLM-5.3 comes with a paid plan. They run on our partner's servers in the USA. What you may do with the answers is still the terms; the model refuses less, the law does not.
| Model | Good for | Attachments |
|---|---|---|
| GLM-5.3 | The newest and strongest of the three: hard reasoning, long documents, a million tokens of context. Carries the green Frontier tag too, and comes with a paid plan. This is what Auto picks; a message with files goes to Claude Opus 5.5 instead, or one with only pictures sometimes to PocketPhantom Noir. Pick GLM-5.3 yourself and PocketPhantom Noir looks at your pictures for it. | Images, through PocketPhantom Noir; PDFs as text |
| GLM-5.2 | The previous large model, with the same million tokens of context. | Images, through PocketPhantom Noir; PDFs as text |
| PocketPhantom Noir | The one that reads pictures: screenshots, photos, scanned images. 256K context. | Images; PDFs as text |
The desktop app is being rebuilt from scratch and the models it will offer are still being decided; see the desktop app.
Effort
Effort is how long the model may think before it answers. The sliders button in the chat bar opens the effort rail; drag it, or use the arrow keys.
| Level | What it means | Use it for |
|---|---|---|
| Low | Quick answers, little deliberation. | Simple questions, lookups, short edits. |
| Medium | Balanced. The default. | Most things. |
| High | Thinks longer before answering. Slower, more thorough. | Planning, debugging, anything with several steps. |
| Max | Thinks as hard as it can. Slowest. | The hard stuff: a proof, a gnarly bug, a long document. |
Higher effort uses more of your monthly AI. The setting is remembered in this browser.
Web search
On unless you switch it off. When an answer needs something current, the model looks it up (a few searches per answer) and tells you where it found it. The Search the web switch in Settings › Search and tools turns it off for your account; every answer then keeps to what the model already knows.
Something missing or wrong? Tell us.