Compare models

Pick up to 4models. Specs render side-by-side. Share the URL — it's stateless.

Comparing agents instead? Switch to agent compare →

Selected

GPT-5 Pro×

Llama 4 Scout×

GPT-5.5×

Gemini 3 Pro×Clear all

	GPT-5 Pro OpenAI	Llama 4 Scout Meta	GPT-5.5 OpenAI	Gemini 3 Pro Google
Vendor	OpenAI	Meta	OpenAI	Google
Family	GPT	Llama	GPT	Gemini
Release date	2025-08-07	2025-04-05	2026-04-23	2026-04-22
Context window	400,000 tokens	10,000,000 tokens	400,000 tokens	2,000,000 tokens
Parameters	—	109B	—	—
Modality	text, vision, audio	text, vision	text, vision, audio	text, vision, audio, video
License	proprietary	Llama 4 Community License	proprietary	proprietary
Source	proprietary	open weights	proprietary	proprietary
Description	Higher-reasoning variant of GPT-5, used for harder scientific and research tasks (e.g., immunology breakthroughs at NIH). Same 400K context as GPT-5; trades latency for depth.	—	OpenAI's smartest and most intuitive model — successor to GPT-5, with major coding, research, and document workflow gains. GPT-5.5 Pro variant also available for heavier reasoning.	Google DeepMind's flagship multimodal model. Best-in-class for multimodal understanding and agentic / vibe coding workflows. A "Deep Think" reasoning mode is available to AI Ultra subscribers.
Links	vendor →	weights →	vendor →	vendor →
Benchmarks
MMLU-Pro	—	—	87.6%	87.1%
GPQA-D	—	—	89.4%	88.6%
HumanEval	—	—	96.8%	96.5%
Aider	—	—	89.7%	84.8%
AIME-25	—	—	94.2%	92.8%
LiveCB	—	—	78.9%	76.1%
MMMU	—	—	78.4%	84.3%