Qwen4-Max benchmark charts compared against Opus5.5, GPT6, and Kimi K3.1

Snoo26837 · reddit · 2026-11-16

A repost of the Qwen4-Max release whose substance is the official benchmark suite: Qwen4-Max versus Qwen3.8-Max-0902, Opus5.5, GPT6.1-Sol, GPT6-Astra, and Kimi K3.1 across DeepSWE, CodeArena, Terminal-Bench 4.0, AA-Briefcase, Agents' Last Exam, JobBench, Automationbench-AA, Humanity's Last Exam, HealthBench Professional and a dozen more, leading or tying on most agentic coding and knowledge-work tasks.

Related event: Qwen4-Max Benchmarks Leak: 50% Fewer Tokens, 60% Cheaper(3 posts)→

Original post →

More from Models

Models channel →