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

DeepSeek-V3.1×

Grok 4×

Llama 4 Maverick×

Gemini 3 Pro×Clear all

	DeepSeek-V3.1 DeepSeek	Grok 4 xAI	Llama 4 Maverick Meta	Gemini 3 Pro Google
Vendor	DeepSeek	xAI	Meta	Google
Family	DeepSeek	Grok	Llama	Gemini
Release date	2025-08-21	2025-07-09	2025-04-05	2026-04-22
Context window	128,000 tokens	256,000 tokens	1,000,000 tokens	2,000,000 tokens
Parameters	671B	—	400B	—
Modality	text	text	text, vision	text, vision, audio, video
License	MIT	proprietary	Llama 4 Community License	proprietary
Source	open weights	proprietary	open weights	proprietary
Description	Large MoE open-weight model. Predecessor to DeepSeek-V4.	—	Mixture-of-experts open-weight model from Meta.	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	weights →	vendor →	weights →	vendor →
Benchmarks
MMLU-Pro	—	82.8%	80.5%	87.1%
GPQA-D	—	87.5%	—	88.6%
HumanEval	—	—	—	96.5%
Aider	—	72.5%	—	84.8%
AIME-25	—	86.1%	—	92.8%
LiveCB	—	65.3%	—	76.1%
MMMU	—	—	68.5%	84.3%