All AI Models Supported — GPT-5, Claude 4.5, Gemini 3 & More

GPT OSS 120B

OpenAI

OpenAI's largest open-source model — same company as ChatGPT. Powerful reasoning and writing.

GPT OSS 20B

OpenAI

OpenAI's compact open-source model — fast responses with solid quality.

Claude Sonnet 4.6

Anthropic

Anthropic's best balance of intelligence and speed. Excellent for analysis, coding, and long documents.

Claude Haiku 4.5

Anthropic

Anthropic's fastest and most compact model — great for high-volume tasks.

Gemini 2.5 Pro

Google

Google's most capable model. State-of-the-art reasoning, coding, and multimodal understanding.

Gemini 2.5 Flash

Google

Google's fast and efficient model — best speed-to-quality ratio in its class.

Gemma 3 12B

Google

Google's open model — strong multilingual reasoning with Gemini DNA.

Llama 4 Scout

Meta

Meta's latest Llama 4 — next-generation MoE architecture, rivals frontier models.

Llama 3.3 70B

Meta

Meta's smartest open model — comparable to GPT-4o quality at zero cost.

Llama 3.1 70B

Meta

Reliable GPT-4-class open model. Great for analysis, coding and long answers.

Llama 3.1 8B

Meta

Lightweight and fast — GPT-3.5 speed with solid quality.

Llama Vision 11B

Meta

Meta's vision-capable model — understands images and text, like GPT-4V.

DeepSeek R1

DeepSeek

The viral reasoning model — shows its thinking step-by-step before answering.

DeepSeek Coder

DeepSeek

DeepSeek's specialized coding model — fast and reliable for programming tasks.

DeepSeek V3

DeepSeek

DeepSeek's flagship chat model — fast, capable, and excellent at math and coding tasks.

Kimi K2.5

Moonshot

China's Kimi model — exceptional at reading very long documents. Like Claude for long context.

Qwen3 Max

Alibaba

Alibaba's latest flagship — powerful and efficient, competitive with top Western models.

QwQ 32B

Alibaba

Alibaba's thinking model — like OpenAI o1, reasons step-by-step before answering.

Qwen Coder 32B

Alibaba

Alibaba's coding specialist — rivals GitHub Copilot for code generation and debugging.

Mistral Large 3

Mistral

Mistral's flagship model — strong reasoning and coding from the European AI leader.

Mistral Small 3.1

Mistral

Fast and capable — beats GPT-3.5, approaches GPT-4 quality at great speed.

Command A

Cohere

Cohere's enterprise model — excellent at structured tasks, retrieval, and business workflows.

Amazon Nova Pro

Amazon

Amazon's latest AI model — strong at multimodal tasks and enterprise workflows.

MiniMax M2.7

MiniMax

MiniMax's powerful model — ultra-long context window and strong Chinese/English bilingual support.

Doubao Seed 2.0

ByteDance

ByteDance's Doubao Seed — competitive reasoning and strong multilingual performance.