gpt-oss-20bvsQwen3.5

gpt-oss-20b
Qwen3.5
Specifications
Parameters
21B
397B
Context window
128k
1M
Benchmarks
Coding
SWE-Bench Verified
60.7%
76.4%Best
Multilingual coding
SWE-Bench Multilingual
69.3%
Agentic terminal coding
Terminal-Bench 2.0
52.5%
Web browsing
BrowseComp
69%
Multidisciplinary reasoning
Humanity's Last Exam · no tools
10.9%
28.7%Best
Multidisciplinary reasoning
Humanity's Last Exam · with tools
17.3%
Science
GPQA Diamond
71.5%
88.4%Best
General knowledge
MMLU
85.3%
Agentic computer use
OSWorld-Verified
62.2%
Chart reasoning
CharXiv Reasoning
80.8%
Multimodal reasoning
MMMU-Pro
79%
Multimodal
MMMU
85%
Overview
CompanyOpenAIQwen
Release dateAug 5 2025Feb 16 2026
AccessOpen WeightOpen Weight

Which is better: gpt-oss-20b or Qwen3.5?

Qwen3.5 leads gpt-oss-20b on 3 of the 3 benchmarks they both report (SWE-Bench Verified, Humanity's Last Exam, GPQA Diamond). gpt-oss-20b shipped 195 days before Qwen3.5, so benchmark comparisons should account for the intervening progress.

gpt-oss-20b has 21B parameters, while Qwen3.5 has 397B. Context windows are 128k (gpt-oss-20b) vs 1M (Qwen3.5).

On SWE-Bench Verified, Qwen3.5 leads at 76.4% vs gpt-oss-20b at 60.7%. On Humanity's Last Exam · no tools, Qwen3.5 leads at 28.7% vs gpt-oss-20b at 10.9%. On GPQA Diamond, Qwen3.5 leads at 88.4% vs gpt-oss-20b at 71.5%.

Frequently asked questions

gpt-oss-20b was released by OpenAI on Aug 5 2025.

Qwen3.5 was released by Qwen on Feb 16 2026.

Qwen3.5 leads on SWE-Bench Verified — gpt-oss-20b 60.7% vs Qwen3.5 76.4%.

Qwen3.5 leads on Humanity's Last Exam · no tools — gpt-oss-20b 10.9% vs Qwen3.5 28.7%.

gpt-oss-20b has a 128k context window; Qwen3.5 has 1M.

Other comparisons