LLMs

OpenAI flagship multimodal model — fast, fluent, and audio-native

OpenAI flagship multimodal model

Anthropic most capable model — deep reasoning, coding, and 200K context

Anthropic flagship with 200K context

Google most capable model — deep reasoning, 1M context, multimodal

OpenAI cheapest model — most of GPT-4 quality at 1/10th cost

Anthropic fastest most affordable model

Google fast multimodal model — 1M token context, real-time inference

xAI frontier model — real-time knowledge, code generation, long context

Meta efficient model — runs on laptops, strong performance

Meta open-weight model — near-frontier performance, self-hostable

Mistral flagship model — European frontier, code generation, reasoning

Mixture-of-experts model — fast inference, strong performance

Alibaba open-weight model — Chinese frontier, code generation, math

Open-weight model rivaling GPT-4 at 1/10th cost

OpenAI reasoning model — chain-of-thought, math, coding

Cohere enterprise model — RAG-optimized, long context, multilingual