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

GPT-5 Pro×

Qwen3.6-27B×

GPT-5.6 Sol×

	GPT-5 Pro OpenAI	Qwen3.6-27B Alibaba	GPT-5.6 Sol OpenAI	o3 OpenAI
Vendor	OpenAI	Alibaba	OpenAI	OpenAI
Family	GPT	Qwen	GPT	o-series
Release date	2025-08-07	2026-04-22	2026-06-26	2025-04-16
Context window	400,000 tokens	262,144 tokens	—	200,000 tokens
Parameters	—	27B (dense)	—	—
Modality	text, vision, audio	text, vision, video	text, vision, audio	text, vision
License	proprietary	Apache-2.0	proprietary	proprietary
Source	proprietary	open weights	proprietary	proprietary
Description	Higher-reasoning variant of GPT-5, used for harder scientific and research tasks (e.g., immunology breakthroughs at NIH). Same 400K context as GPT-5; trades latency for depth.	Alibaba's first dense open-weight in the 3.6 family. Strong agentic-coding scores (77.2 SWE-bench Verified, matching Claude 4.5 Opus on Terminal-Bench 2.0). Supports 201 languages and multimodal text/image/video input.	Flagship of OpenAI's GPT-5.6 three-model suite (Sol / Terra / Luna), released as a limited preview. Strong gains in coding, cybersecurity, biology, and long-horizon agentic work. Pricing: $5 / $30 per million input/output tokens.	Reasoning-focused model in the o-series.
Links	vendor →	weights →	vendor →
Benchmarks
MMLU-Pro	—	78.9%	—	—
GPQA-D	—	—	—	87.7%
HumanEval	—	91.5%	—	—
Aider	—	70.2%	—	—
AIME-25	—	—	—	88.9%
MMMU	—	70.3%	—	—