Compare AI models

Specifications
Parameters
—27B
Context window
1M256k
Benchmarks
SWE-Bench Verified
93.9%77.2%
SWE-Bench Multilingual
87.3%71.3%
Terminal-Bench 2.0
82%59.3%
Humanity's Last Exam · no tools
56.8%24%
GPQA Diamond
94.5%87.8%
CharXiv Reasoning
86.1%78.4%
Overview
CompanyAnthropicQwen
Release dateApr 7 2026Apr 22 2026
AccessClosedOpen Weight
Model detailsView modelView model

Frequently asked questions

Claude Mythos Preview leads Qwen3.6-27B on 6 of the 6 benchmarks they both report. Claude Mythos Preview shipped 15 days before Qwen3.6-27B, so benchmark comparisons should account for the intervening progress.

Context windows are 1M (Claude Mythos Preview) vs 256k (Qwen3.6-27B). Claude Mythos Preview is closed, while Qwen3.6-27B is open weight.

On SWE-Bench Verified, Claude Mythos Preview leads at 93.9% vs Qwen3.6-27B at 77.2%. On SWE-Bench Multilingual, Claude Mythos Preview leads at 87.3% vs Qwen3.6-27B at 71.3%. On Terminal-Bench 2.0, Claude Mythos Preview leads at 82% vs Qwen3.6-27B at 59.3%. On Humanity's Last Exam · no tools, Claude Mythos Preview leads at 56.8% vs Qwen3.6-27B at 24%. On GPQA Diamond, Claude Mythos Preview leads at 94.5% vs Qwen3.6-27B at 87.8%. On CharXiv Reasoning, Claude Mythos Preview leads at 86.1% vs Qwen3.6-27B at 78.4%.