gpt-oss-20b
OpenAI · 2 thinking configurations
Intelligence9.0
Input / 1M$0.07
Output / 1M$0.18
Context131.1K
Tariff: OpenAI · Source: artificial-analysis · USD per million tokens
Thinking configuration
What does more thinking buy you?
Compare this model’s measured thinking configurations.
Scroll the chart ↔ · tap a point for details
Efficient frontier: best measured score at each cost1 measured configurations · 1 without both values
Selected: thinking high · Click a point to switch configuration
gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for...
Intelligence IndexArtificial Analysis composite index, not a percentage. Different versions of the index are not directly comparable.
9.0 pointsGPQA DiamondGraduate-level scientific questions. Measures specialist reasoning rather than web-search quality.
68.8 %Humanity’s Last ExamDifficult questions across many disciplines, compared under the same evaluation setup.
11.0 %SciCodeCode generation for scientific research problems.
38.9 %Terminal-Bench 2.1Agent tasks in a terminal environment. Results also depend on the evaluation harness.
13.9 %τ²-BenchTool use and interaction in assistance scenarios. Does not replace an evaluation of your integration.
60.2 %IFBenchVerifiable instruction following. A proxy for precision rather than a benchmark of creative style.
65.1 %Omniscience accuracyFactual accuracy in Artificial Analysis's Omniscience evaluation.
16.0 %Specifications from the catalog
Inputtext
Outputtext
Tool callingYes
Structured outputYes
Speed165.4 tok/s
First token0.81 s
Maximum output32.8K
Catalog listing date2025-08-05