All AI models · LLM benchmarks · Methodology

OpenAI

GPT OSS 20B

Open-weight GPT model for self-hosted reasoning and instruction-following workloads

Model facts

Context window
131K tokens
Maximum output
33K tokens
Cheapest paid input
$0.02 per 1M tokens
Cheapest paid output
$0.1 per 1M tokens
Open weights
yes
Providers
51

Capabilities and modalities

reasoning, tool calling, structured output, attachments, temperature control, open weights, text, image.

Local hardware estimate

At 4K context: Q4_K_M 13.8 GiB, Q8_0 23.5 GiB, F16 41.8 GiB working memory.

Parameters
20.9 billion
Architecture
gptoss
Model license
apache-2.0

Planning estimate adapted from the MIT-licensed whichllm estimator using metadata from Hugging Face; not a vendor minimum requirement.

Artificial Analysis benchmarks

Compare this model on the full LLM benchmark leaderboard.

ConfigurationAAICodingMathOutput tokens/s
high15.220.789.3127.7
low14.4not measured62.3146.1

Observed price history

  • 2026-08-24: $0.03 input / $0.13 output per 1M tokens
  • 2026-08-25: $0.02 input / $0.1 output per 1M tokens

API providers