BlogLLM Benchmark Scores as Published, Side by Side

gpt-oss-20b: Benchmark Scores as Published

Every result the catalog holds for gpt-oss-20b: 29 tests in 9 areas, from 2 pages. Each was printed by the company named beside it. Each number is as printed; nothing is ranked or rescored.

Developer OpenAIReleased 2025-08-05Context window 131KMax output 131KInput TextOutput TextWeights Open (Apache 2.0)Data version 2026103001, as of 2026-10-30

Agentic Coding

Science and Math

Long Context

TestValuePrinted bySetup as printed
AA-LCR32.88%NVIDIA, NVIDIA Nemotron 3.5 Lightning 30B-A3B model cardchecked 2026-10-19; printed on 1 more page-

General Knowledge and Reasoning

Code Generation

TestValuePrinted bySetup as printed
LiveCodeBench v661%NVIDIA, NVIDIA Nemotron 3 Nano 30B-A3B model cardchecked 2026-10-15-

Tool Use and Workflows

Instruction Following

TestValuePrinted bySetup as printed
IFBench68.5NVIDIA, NVIDIA Nemotron 3.5 Lightning 30B-A3B model cardchecked 2026-10-19; printed on 1 more page-
MultiChallenge33.8%NVIDIA, NVIDIA Nemotron 3 Nano 30B-A3B model cardchecked 2026-10-15-

Multilingual

Preference and Composite Rankings

More from OpenAI

Every number here is read from the same catalog file as the benchmark table, data version 2026103001. Where one page printed the model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.