OpenAI

gpt-oss-20B (low)

Unknown Size

By OpenAI • Released 2025-08-05

Capability Radar

Avg Score
40

Across all benchmarks

Participated
12
Benchmarks

Benchmark Performance

Benchmark Category Score
MMLU-Pro
Knowledge
71.8
LiveCodeBench
Coding
65.2
AIME 2025
Reasoning
62.3
GPQA Diamond
Knowledge
61.1
IFBench
Agent
57.8
𝜏²-Bench Telecom
Reasoning Knowledge
50.3
SciCode
Reasoning Knowledge
34
LCR
Long-Context Reasoning
31
Artificial Analysis Intelligence Index
Knowledge
20.8
Artificial Analysis Coding Index
Coding
14.4
HLE
Knowledge Multi-Modal
5.1
Terminal-Bench Hard
Agent Coding
4.5