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

Gemini 3 Pro×

Grok 4×

DeepSeek-V4-Pro×

DeepSeek-R1×Clear all

	Gemini 3 Pro Google	Grok 4 xAI	DeepSeek-V4-Pro DeepSeek	DeepSeek-R1 DeepSeek
Vendor	Google	xAI	DeepSeek	DeepSeek
Family	Gemini	Grok	DeepSeek	DeepSeek
Release date	2026-04-22	2025-07-09	2026-04-22	2025-01-20
Context window	2,000,000 tokens	256,000 tokens	1,000,000 tokens	128,000 tokens
Parameters	—	—	1.6T (49B active)	671B
Modality	text, vision, audio, video	text	text	text
License	proprietary	proprietary	MIT	MIT
Source	proprietary	proprietary	open weights	open weights
Description	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.	—	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.	Reasoning-focused open-weight model.
Links	vendor →	vendor →	weights →	weights →
Benchmarks
MMLU-Pro	87.1%	82.8%	84.2%	—
GPQA-D	88.6%	87.5%	82.4%	—
HumanEval	96.5%	—	95.1%	—
Aider	84.8%	72.5%	80.1%	—
AIME-25	92.8%	86.1%	88.6%	79.8%
LiveCB	76.1%	65.3%	72.4%	—
MMMU	84.3%	—	—	—