Back to all LLMs

GPT-OSS 120B

131K context with Apache 2.0 license for full customization and self-hosting. Configurable reasoning depth with <think> tags and single 80GB GPU deployment for self-hosted RAG. If you want to compare the best LLMs for your data, try Agentset.

Leaderboard Rank
#16
of 16
ELO Rating
1242
#16
Win Rate
14.2%
#16
Latency
11302ms
#8

Model Information

Provider
OpenAI
License
Open Source
Input Price per 1M
$0.04
Output Price per 1M
$0.19
Context Window
131K
Release Date
2025-08-05
Model Name
gpt-oss-120b
Total Evaluations
1350

Performance Record

Wins192 (14.2%)
Losses1029 (76.2%)
Ties129 (9.6%)
Wins
Losses
Ties

LLMs Are Just One Piece of RAG

Agentset gives you a managed RAG pipeline with the top-ranked models and best practices baked in. No infrastructure to maintain, no LLM orchestration to manage.

Trusted by teams building production RAG applications

5M+
Documents
1,500+
Teams
99.9%
Uptime

Performance Overview

ELO ratings by dataset

GPT-OSS 120B's ELO performance varies across different benchmark datasets, showing its strengths in specific domains.

GPT-OSS 120B - ELO by Dataset

Detailed Metrics

Dataset breakdown

Performance metrics across different benchmark datasets, including accuracy and latency percentiles.

PG

ELO 128223.8% WR107W-332L-11T

Quality Metrics

Correctness
4.87
Faithfulness
4.87
Grounding
4.87
Relevance
4.90
Completeness
4.83
Overall
4.87

Latency Distribution

Mean
19128ms
Min
1317ms
Max
69491ms

MSMARCO

ELO 125413.3% WR60W-340L-50T

Quality Metrics

Correctness
4.93
Faithfulness
4.93
Grounding
4.93
Relevance
4.97
Completeness
4.80
Overall
4.91

Latency Distribution

Mean
5616ms
Min
1255ms
Max
20330ms

SciFact

ELO 11905.6% WR25W-357L-68T

Quality Metrics

Correctness
4.70
Faithfulness
4.80
Grounding
4.80
Relevance
4.73
Completeness
4.57
Overall
4.72

Latency Distribution

Mean
9160ms
Min
1606ms
Max
35709ms

Build RAG in Minutes, Not Months

Agentset gives you a complete RAG API with top-ranked LLMs and smart retrieval built in. Upload your data, call the API, and get grounded answers from day one.

import { Agentset } from "agentset";

const agentset = new Agentset();
const ns = agentset.namespace("ns_1234");

const results = await ns.search(
  "What is multi-head attention?"
);

for (const result of results) {
  console.log(result.text);
}

Compare Models

See how it stacks up

Compare GPT-OSS 120B with other top llms to understand the differences in performance, accuracy, and latency.