Labs

ChatGPT's parent. The GPT-5.6 line is a three-rung ladder: Sol for hard work, Terra for most production traffic, Luna for volume.

Claude. Opus 5 leads public intelligence rankings. Fable 5 is the named flagship and the expensive one. Sonnet 5 is the daily driver.

Gemini is natively multimodal. 3.7 Flash is the current pick for mixed media and long documents. 3.5 Pro is still slipping.

Grok. Wired into X and the live web. Grok 4.6 sits on the frontier scoreboard at a mid-tier price. Context is 500K, half of the pack.

Open weights under MIT. V4 Pro is the first downloadable model that actually sits in the frontier cluster. Flash is the volume lane.

Two flagships: Qwen3.8-Max is a proprietary multimodal API. The 2.4T checkpoint is downloadable, text-only, and commercially restricted.

Kimi. Built for long agentic coding runs. K3 is the strongest model you can download today. The license is Moonshot's own, not MIT.

GLM. 5.3 matches Kimi on the composite at a lower API price. Weights are not public, so this is still a rental.

Muse is closed and powers Meta AI across Facebook, Instagram, and WhatsApp. Llama 4 still ships for local work. Glimmer is the small Apache agent.

Europe's lab. Smaller, cheaper models with EU data residency. Large 3 is the hosted flagship. Ministral is the on-device one.

Nova on Bedrock. Not a frontier intelligence pick. Useful if you already live in AWS and want vision plus tools without a second vendor.

Command A is built for RAG and enterprise search, not olympiad puzzles. Strong when the answer has to come from your documents.

Nemotron publishes weights, data, and the training recipe. 3.5 Lightning is a 30B local agent, not a board leader. It is the most open release here.

M3 is a cheap multimodal all-rounder with a million-token window. Weights exist. The community license restricts commercial use.