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 Scout×

DeepSeek-V4-Pro×

Mistral Large 2×

Llama 4 Maverick×Clear all

	Llama 4 Scout Meta	DeepSeek-V4-Pro DeepSeek	Mistral Large 2 Mistral	Llama 4 Maverick Meta
Vendor	Meta	DeepSeek	Mistral	Meta
Family	Llama	DeepSeek	Mistral	Llama
Release date	2025-04-05	2026-04-22	2024-07-24	2025-04-05
Context window	10,000,000 tokens	1,000,000 tokens	128,000 tokens	1,000,000 tokens
Parameters	109B	1.6T (49B active)	123B	400B
Modality	text, vision	text	text	text, vision
License	Llama 4 Community License	MIT	Mistral Research License	Llama 4 Community License
Source	open weights	open weights	open weights	open weights
Description	—	DeepSeek's flagship open-weight MoE. 1.6T parameters with 49B activated, 1M-token context, and a hybrid attention scheme (CSA + HCA) that delivers long-context inference at ~27% of V3.2's FLOPs.	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.	Mixture-of-experts open-weight model from Meta.
Links	weights →	weights →	vendor →weights →	weights →
Benchmarks
MMLU-Pro	—	84.2%	—	80.5%
GPQA-D	—	82.4%	—	—
HumanEval	—	95.1%	89.0%	—
Aider	—	80.1%	—	—
AIME-25	—	88.6%	—	—
LiveCB	—	72.4%	—	—
MMMU	—	—	—	68.5%