Replying to Avatar Maple AI

We need your help to shape Maple 2.0!

Join our beta for Maple 1.99 to test the foundation for 2.0!

• Background the app & refresh without losing replies

• Stop a chat while it's processing

• Better reliability

Help us build a better Maple!

https://testflight.apple.com/join/zjgtyAeD

Testing Notes:

• Old chats move to "Archived" (read-only) folder

• New chats won’t sync to web…yet

• Testing first on iOS

• You can go back anytime: Uninstall TestFlight & Install App Store version (old chats editable again!)

What LLM models does maple offer?

Reply to this note

Please Login to reply.

Discussion

We offer seven models at the moment, each running in a secure enclave for end-to-end encryption:

DeepSeek R1 671B

OpenAI GPT-OSS 120B

Qwen3 Coder 480B

Meta Llama 3.3 70B

Google Gemma 3 27B

Mistral Small 3.1 24B

Qwen 2.5 72B

More details on our blog: https://blog.trymaple.ai/maple-ai-model-guide-with-example-prompts/

Awesome, thanks for the info and the link. How do you feel these compare to the current commercial models? I've been using Claude Sonnet 4.5 for coding and it is generally considered the best.

Claude has become the current standard for coding. Qwen3 Coder performs well on benchmarks. It’s also less expensive than Claude and comes with the benefit of data privacy.

With most models, benchmarks only get you so far. A lot of it is personal preference and situational context. You have to plug them into your workflow to see how they perform for you.