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

Llama 4 Maverick×

GPT-5.5×

Mistral Large 2×

DeepSeek-V3.1×Clear all

	Llama 4 Maverick Meta	GPT-5.5 OpenAI	Mistral Large 2 Mistral	DeepSeek-V3.1 DeepSeek
Vendor	Meta	OpenAI	Mistral	DeepSeek
Family	Llama	GPT	Mistral	DeepSeek
Release date	2025-04-05	2026-04-23	2024-07-24	2025-08-21
Context window	1,000,000 tokens	400,000 tokens	128,000 tokens	128,000 tokens
Parameters	400B	—	123B	671B
Modality	text, vision	text, vision, audio	text	text
License	Llama 4 Community License	proprietary	Mistral Research License	MIT
Source	open weights	proprietary	open weights	open weights
Description	Mixture-of-experts open-weight model from Meta.	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.	Large text-only Mistral model with a 128K context window and 123B parameters, tuned for strong instruction following and long-context reasoning. Mistral's flagship open-weights release in the Large 2 line.	Large MoE open-weight model. Predecessor to DeepSeek-V4.
Links	weights →	vendor →	vendor →weights →	weights →
Benchmarks
MMLU-Pro	80.5%	87.6%	—	—
GPQA-D	—	89.4%	—	—
HumanEval	—	96.8%	89.0%	—
Aider	—	89.7%	—	—
AIME-25	—	94.2%	—	—
LiveCB	—	78.9%	—	—
MMMU	68.5%	78.4%	—	—