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-R1×

DeepSeek-V4-Pro×

Llama 4 Maverick×

Gemini 2.5 Flash×Clear all

	DeepSeek-R1 DeepSeek	DeepSeek-V4-Pro DeepSeek	Llama 4 Maverick Meta	Gemini 2.5 Flash Google
Vendor	DeepSeek	DeepSeek	Meta	Google
Family	DeepSeek	DeepSeek	Llama	Gemini
Release date	2025-01-20	2026-04-22	2025-04-05	2025-06-17
Context window	128,000 tokens	1,000,000 tokens	1,000,000 tokens	1,000,000 tokens
Parameters	671B	1.6T (49B active)	400B	—
Modality	text	text	text, vision	text, vision, audio
License	MIT	MIT	Llama 4 Community License	proprietary
Source	open weights	open weights	open weights	proprietary
Description	Reasoning-focused open-weight model.	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.	Mixture-of-experts open-weight model from Meta.	Google's cost-and-latency-optimized Gemini 2.5 variant. Generally available June 2025 — designed for high-throughput agentic workflows with sub-second first-token latency.
Links	weights →	weights →	weights →
Benchmarks
MMLU-Pro	—	84.2%	80.5%	—
GPQA-D	—	82.4%	—	—
HumanEval	—	95.1%	—	—
Aider	—	80.1%	—	—
AIME-25	79.8%	88.6%	—	—
LiveCB	—	72.4%	—	—
MMMU	—	—	68.5%	—