Compare AI models

Specifications
Parameters
—27B/35B/122B
Context window
400k256k
API pricing
Input price
$0.75—
Output price
$4.50—
Cached input price
$0.075—
Cheapest input
$0.375OpenAI—
Cheapest output
$2.25OpenAI—
$375OpenAI list price—
Benchmarks
Terminal-Bench 2.0
60%49.4%
Humanity's Last Exam · no tools
28.2%25.3%
Humanity's Last Exam · with tools
41.5%47.5%
GPQA Diamond
88%86.6%
OSWorld-Verified
72.1%58%
MMMU-Pro
76.6%76.9%
Overview
CompanyOpenAIQwen
Release dateMar 17 2026Feb 24 2026
AccessClosedOpen Weight
Model detailsView modelView model

Frequently asked questions

GPT-5.4 mini leads Qwen3.5 Medium on 4 of the 6 benchmarks they both report. Only GPT-5.4 mini has a verified first-party price for an available API: $0.75 per million input tokens and $4.50 per million output tokens. No currently available pay-as-you-go API rate is tracked for Qwen3.5 Medium. Qwen3.5 Medium shipped 21 days before GPT-5.4 mini, so benchmark comparisons should account for the intervening progress.

Context windows are 400k (GPT-5.4 mini) vs 256k (Qwen3.5 Medium). GPT-5.4 mini is closed, while Qwen3.5 Medium is open weight.

On Terminal-Bench 2.0, GPT-5.4 mini leads at 60% vs Qwen3.5 Medium at 49.4%. On Humanity's Last Exam · no tools, GPT-5.4 mini leads at 28.2% vs Qwen3.5 Medium at 25.3%. On Humanity's Last Exam · with tools, Qwen3.5 Medium leads at 47.5% vs GPT-5.4 mini at 41.5%. On GPQA Diamond, GPT-5.4 mini leads at 88% vs Qwen3.5 Medium at 86.6%. On OSWorld-Verified, GPT-5.4 mini leads at 72.1% vs Qwen3.5 Medium at 58%. On MMMU-Pro, Qwen3.5 Medium leads at 76.9% vs GPT-5.4 mini at 76.6%.