TransistorOSDocs
ResearchDocs
Connect

Supported Models

Transistor works with the leading AI providers. Connect with your own API key and switch models at any time without losing your memory, skills, or history.

Top models by Arena ranking

Rankings are sourced from the Arena leaderboard at arena.ai, which scores models based on blind head-to-head user evaluations. The list below reflects the current top performers across providers.

RankModelProvider
1Claude Opus 4.6 (Thinking)Anthropic
2Claude Opus 4.6Anthropic
3Gemini 3.1 ProGoogle
4Grok 4xAI
5Gemini 3 ProGoogle
6GPT-5OpenAI
7Gemini 3 FlashGoogle
8Claude Sonnet 4.6Anthropic
9Qwen 3.5 MaxAlibaba
Source: arena.ai leaderboard

All supported providers

ProviderNotes
AnthropicTop-ranked models overall. Recommended as a starting point for most users.
OpenAIStrong general capability. GPT-5 competitive with top Anthropic models.
GoogleMultiple price and speed tiers. Gemini 3 Flash offers strong performance at low cost.
xAIGrok 4 competitive with top models. Good for real-time information access.
AlibabaQwen 3.5 Max is highly competitive and often lower cost than Western alternatives.
CustomAny OpenAI-compatible API. Works with Ollama, LM Studio, Together AI, Fireworks, and others.

Switching models

You can switch models at any time from the model picker in the chat interface. Your memory, skills, and conversation history are stored by Transistor, not by the model provider, so switching has no effect on your agent's accumulated context.

Using multiple models

Different conversations can use different models. You might use a fast, inexpensive model for quick lookups and a more capable model for complex research. You can set a default model in Settings and override it per-conversation.

Bring your own key

Transistor does not proxy your requests through its own servers. Your API key is stored locally and your conversations go directly from your device to the model provider. Transistor never sees the content of your conversations with the AI.