🧠 AI Models
22 models across 13 providers
Llama 3.3 70B
NVIDIA NIM
Meta's latest 70B model. Great all-rounder for agents and chat.
Llama 3.3 70B
Groq
Meta's Llama 3.3 on Groq LPU — ultra fast inference.
Mixtral 8x22B
OpenRouter
Mistral's MoE model. Top-tier reasoning at reasonable cost.
Mixtral 8x7B
Groq
Fast MoE model on Groq. Great for quick completions.
Llama 3 70B
Together AI
Llama 3 70B at competitive prices on Together AI.
DeepSeek R1
Together AI
DeepSeek's reasoning model. Great for complex math/code.
DeepSeek V3
DeepSeek
Extremely cost-effective 671B MoE model.
Gemini 2.5 Pro
Google Gemini
Google's most capable model. 1M context, multimodal.
Gemini 2.0 Flash
Google Gemini
Fast and cheap. Great for high-throughput use cases.
Claude Sonnet 4
Anthropic
Best balance of speed and capability. Excellent for coding.
Claude Haiku 4
Anthropic
Fastest Claude. Great for high-throughput tasks.
GPT-4.1
OpenAI
Latest GPT with 1M context. Excellent for coding agents.
Nemotron
NVIDIA NIM
NVIDIA's in-house model. Great for enterprise use.
Command R+
Cohere
Enterprise-grade with strong RAG capabilities.
Mistral Large
Mistral
Mistral's flagship. Top-tier reasoning in multiple languages.
Codestral
Mistral
Mistral's code-specialized model. Great fill-in-the-middle.
Qwen 2.5 72B
Together AI
Alibaba's Qwen 2.5. Strong multilingual performance.
Llama 3.1 Nemotron 70B
NVIDIA NIM
NVIDIA's fine-tuned Llama. Enterprise-optimized.