Compare AI models

Specifications
Parameters
—27B
Context window
400k256k
API pricing
Input price
$1.75—
Output price
$14.00—
Cached input price
$0.175—
Cheapest input
$0.875OpenAI—
Cheapest output
$7.00OpenAI—
$1,050OpenAI list price—
Benchmarks
SWE-Bench Verified
80%77.2%
Humanity's Last Exam · no tools
34.5%24%
GPQA Diamond
92.4%87.8%
CharXiv Reasoning
82.1%78.4%
MMMU-Pro
79.5%75.8%
Overview
CompanyOpenAIQwen
Release dateDec 11 2025Apr 22 2026
AccessClosedOpen Weight
Model detailsView modelView model

Frequently asked questions

GPT-5.2 leads Qwen3.6-27B on 5 of the 5 benchmarks they both report. Only GPT-5.2 has a verified first-party price for an available API: $1.75 per million input tokens and $14.00 per million output tokens. No currently available pay-as-you-go API rate is tracked for Qwen3.6-27B. GPT-5.2 shipped 132 days before Qwen3.6-27B, so benchmark comparisons should account for the intervening progress.

Context windows are 400k (GPT-5.2) vs 256k (Qwen3.6-27B). GPT-5.2 is closed, while Qwen3.6-27B is open weight.

On SWE-Bench Verified, GPT-5.2 leads at 80% vs Qwen3.6-27B at 77.2%. On Humanity's Last Exam · no tools, GPT-5.2 leads at 34.5% vs Qwen3.6-27B at 24%. On GPQA Diamond, GPT-5.2 leads at 92.4% vs Qwen3.6-27B at 87.8%. On CharXiv Reasoning, GPT-5.2 leads at 82.1% vs Qwen3.6-27B at 78.4%. On MMMU-Pro, GPT-5.2 leads at 79.5% vs Qwen3.6-27B at 75.8%.