Six new models are in the model list today, three for Claude and three for ChatGPT. Your project keeps the model it already has, and nothing changes until you pick a new one.
Your Bunny runs on the AI platform and the API key you connected to your project. The model you choose there does both jobs: it reads your files when they are indexed, and it answers you in chat. Both platforms released a new generation over the last five weeks, and all of it is selectable now.
Claude
| Model | What Anthropic builds it for | Input / output |
|---|---|---|
| Claude Sonnet 5.5 | The best combination of speed and intelligence | $2 / $10 |
| Claude Opus 5.5 | Long-running agentic coding and knowledge work | $4 / $20 |
| Claude Fable 5.1 | Demanding reasoning and long-horizon agentic work | $10 / $50 |
Sonnet 5.5 is the fastest of the three and costs the least. Fable 5.1 is the most capable and the slowest. Anthropic's own advice is to start with Opus 5.5 for most work, and to move up to Fable 5.1 when that falls short.
All three take up to 1,000,000 tokens in a single request, the same as the Claude models you already had.
ChatGPT
| Model | What OpenAI builds it for | Input / output |
|---|---|---|
GPT-6 Luna, gpt-6-luna |
Focused, high-volume tasks. Its most efficient model | $0.10 / $0.50 |
GPT-6 Sol, gpt-6-sol |
Complex coding and agentic workflows | $2 / $10 |
GPT-6 Astra, gpt-6-astra |
The hardest end-to-end work. Its most capable model | $10 / $50 |
The ChatGPT list shows each model by its id, so look for the name in the second form. Sol and Luna come in at half the price of the GPT-5.6 models of the same name, or less.
All three take up to 1,050,000 tokens in a single request, the same as GPT-5.6.
About the prices
The prices are each platform's base list price per million tokens, in US dollars, on September 29, 2026. Very long requests and cached input are priced differently, so read the platform's pricing page before you decide.
The bill for tokens comes from the platform, on your own API key.
How to switch
- In the chat, the model name at the top is a menu. Pick a model and your next message uses it.
- In Settings, under Query API, pick it in the Model row.
Either way the choice is saved to the project. It uses the key you already saved for that platform, so there is no new key to add. A model works as long as your account with the platform has access to it.
What stays the same
- Your project keeps its model. Nothing was switched for you.
- The models you had are still in the list, from Claude Opus 4.8 to GPT-5.6.
- A new project starts where it did before, on Claude Sonnet 5 or GPT-5.6 Luna, until you pick something else.
- Files that are already indexed stay as they are. A new model reads the files you add after the switch.
More news coming up!
Subscribe to the newsletter from your account profile page so you do not miss anything, and past articles are always at bunnyquery.com/articles.
Go try one. Enjoy your Bunny!
Baksa, Creator of BunnyQuery