Choosing a chat model
What each model is best at, and when the extra cost of a frontier model is worth it.
GLM 5.2 is the default and handles most conversations and builds beautifully at the best value. Reach for a frontier model when the work is genuinely demanding — precise reasoning, careful code, photoreal 3D, or long documents.
The text & chat lineup
| Model | Best for | Tier |
|---|---|---|
| GLM 5.2 | Everyday chat and building — fast, capable, the default for most work. | Budget |
| Opus 4.8 (Anthropic) | Frontier work: photoreal 3D, intricate polish, hard multi-step reasoning. | Premium |
| ParleyX (GPT-5.5) | Top-tier reasoning and long-context analysis for genuinely hard problems. | Premium |
| ParleyX Pro (GPT-5.5 Pro) | Maximum intelligence for one-shot deep analysis. Slower, pricier. | Premium |
| Opus 4.5 (Anthropic) | Nuanced writing, careful code review, long agent tool loops. | Premium |
| Gemini 3.1 Pro (Google) | A 1M-token window — feed it an entire codebase, book, or long PDF. | Standard |
| Venice — Uncensored | Free, unrestricted creative writing and roleplay. Text only. | Free |
Switching models
Open the model picker next to the prompt box and choose. The choice applies to your next turn; you can change it mid-conversation without losing the thread.