OpenAI: gpt-oss-20b
openai/gpt-oss-20b
136th smartest of 182Cheaper than 98% of paidReasoningToolsJSON
Intelligence
14.9
136th of 182
Design Elo
963
Data Viz
Speed
—
tokens/sec
Latency
—
first token
Input price
$0.030
8th cheapest
Context
131K
131K max out
How it compares
Smarter than25%
of all ranked models
Cheaper than98%
of all ranked models
Overview
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...
Benchmarks
independent · Artificial Analysis & Design ArenaArtificial Analysis50th percentile
Intelligence Index
14.9
Coding Index
20.7
GPQA Diamond
69%
Humanity's Last Exam
10%
SciCode
34%
Tau²-Bench (agentic)
60%
Design Arena · Elo78 tournaments
Data Viz
963
Website
877
Providers & pricing (12)
| Provider | In $/M | Out $/M | Context | Uptime |
|---|---|---|---|---|
| CoreWeavefp4 | $0.030 | $0.130 | 131K | 99.9% |
| DeepInfrabf16 | $0.030 | $0.140 | 131K | 99.9% |
| Parasailfp4 | $0.030 | $0.150 | 131K | 78.8% |
| Novitafp4 | $0.040 | $0.150 | 131K | 100% |
| Phala | $0.040 | $0.150 | 131K | 99.9% |
| SiliconFlowfp8 | $0.040 | $0.180 | 131K | 93.9% |
| Together | $0.050 | $0.200 | 131K | — |
| Amazon Bedrock | $0.070 | $0.150 | 131K | — |
| Amazon Bedrock | $0.070 | $0.150 | 131K | 99.9% |
| $0.070 | $0.250 | 131K | 73.1% | |
| Fireworks | $0.070 | $0.300 | 131K | — |
| Groq | $0.075 | $0.300 | 131K | 99.3% |
Specifications
Context window131K
Max output131K
Knowledge cutoffJun 2024
Input modalitiestext
Output modalitiestext
Prompt caching—
Cache read price$0.030/M
ModeratedNo
Open weightsopenai/gpt-oss-20b ↗
gpt-oss-20b FAQ
How much does gpt-oss-20b cost?
gpt-oss-20b costs $0.030 per million input tokens and $0.130 per million output tokens via OpenRouter, making it 8th cheapest of 332 paid models.
How smart is gpt-oss-20b?
gpt-oss-20b scores 14.9 on the Artificial Analysis Intelligence Index, ranking 136th of 182 benchmarked models, with a GPQA Diamond score of 69%.
What is gpt-oss-20b's context window?
gpt-oss-20b supports a 131K-token context window and can output up to 131K tokens. It accepts text input.
Similar models
All gpt-oss-20b alternatives →Compare head-to-head
OpenAI: gpt-oss-20b vs NVIDIA: Nemotron 3 Nano Omni (free)OpenAI: gpt-oss-20b vs OpenAI: gpt-oss-20b (free)OpenAI: gpt-oss-20b vs OpenAI: GPT-4.1 MiniOpenAI: gpt-oss-20b vs Mistral: Mistral Medium 3.1OpenAI: gpt-oss-20b vs Qwen: Qwen3 VL 235B A22B InstructOpenAI: gpt-oss-20b vs Meta: Llama 4 Maverick