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

Grok 4×

Gemini 3.5×

Qwen3.7-Max×

Mistral Large 2×Clear all

	Grok 4 xAI	Gemini 3.5 Google	Qwen3.7-Max Alibaba	Mistral Large 2 Mistral
Vendor	xAI	Google	Alibaba	Mistral
Family	Grok	Gemini	Qwen	Mistral
Release date	2025-07-09	2026-05-20	2026-05-20	2024-07-24
Context window	256,000 tokens	2,000,000 tokens	1,000,000 tokens	128,000 tokens
Parameters	—	—	—	123B
Modality	text	text, vision, audio, video	text	text
License	proprietary	proprietary	proprietary	Mistral Research License
Source	proprietary	proprietary	proprietary	open weights
Description	—	Google DeepMind's next-gen Gemini — positioned as "frontier intelligence with action". Built for complex agentic workflows. Announced at Google I/O 2026.	Alibaba's flagship agent model — 1M-token context, extended-thinking mode, 56.6 on the Artificial Analysis Intelligence Index v4.0 (5th overall, #1 Chinese). 50.8% on Terminal-Bench Hard. Designed for long-horizon agent workloads (hundreds-to-thousands of steps). Closed-weight, $2.50/$7.50 per 1M tokens.	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.
Links	vendor →	vendor →	vendor →	vendor →weights →
Benchmarks
MMLU-Pro	82.8%	88.9%	83.7%	—
GPQA-D	87.5%	91.2%	83.0%	—
HumanEval	—	—	93.9%	89.0%
Aider	72.5%	—	78.4%	—
AIME-25	86.1%	96.1%	90.4%	—
LiveCB	65.3%	—	71.0%	—
MMMU	—	86.0%	—	—