Evalry

gpt-oss-120b

gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases. It activates 5.1B parameters per forward pass and is optimized...

by OpenAI

Overview

Quick stats across all benchmark runs.

Score

34%

279 benchmarks

Avg Latency

28.7s

758 requests

Pricing

Free in / Free out

per 1M tokens

Context

131K

tokens

Alternatives

Models with similar or better quality but different tradeoffs

Same Quality, Cheaper

Models with similar or better performance at a lower cost per token.

ModelCost
Llama 4 Scout-67%
Ministral 3 8B 2512-63%
gpt-oss-20b-48%
Llama 3.3 70B Instruct-45%
Ministral 3 14B 2512-40%

Same Quality, Faster

Models with similar or better performance but lower latency.

ModelLatency
Claude 3 Haiku-87%
Gemini 2.0 Flash-85%
Gemini 3 Flash Preview-80%
Llama 4 Maverick-80%
Gemini 2.5 Flash Lite-79%

Same Cost, Better

Models at a similar price point with higher benchmark scores.

ModelScore
DeepSeek V3.2+24%
Gemini 2.0 Flash+22%
Llama 4 Maverick+20%
Llama 3.3 70B Instruct+13%
Gemini 2.5 Flash Lite+6%

Benchmark Performance

How this model performs across different benchmarks

BenchmarkScoreRank
Venture Capital Terms Benchmark
100%
1 / 21
Spatial Reasoning: Germany
93%
12 / 29
Character Frequency Bench
91%
7 / 29
Money Boy Cultural Literacy Test
46%
42 / 107
Explain Like I'm 5
39%
30 / 47
Categorization Bench
33%
42 / 54
German Memelord Bench
31%
17 / 29
Niederstetten Benchmark
20%
27 / 35

Price vs Performance

Compare cost efficiency across all models

Current model (baseline)
Other models (relative score)
Y-axis shows score difference from shared benchmarks. X-axis uses log scale.

Score Over Time

Performance trends across all benchmark runs

Benchmark Activity

Number of benchmark runs over time

Quickstart

View on OpenRouter

Get started with this model using OpenRouter

import { OpenRouter } from "@openrouter/sdk";

const openrouter = new OpenRouter({
  apiKey: "<OPENROUTER_API_KEY>"
});

const completion = await openrouter.chat.completions.create({
  model: "openai/gpt-oss-120b:free",
  messages: [
    {
      role: "user",
      content: "Hello!"
    }
  ]
});

console.log(completion.choices[0].message.content);

Get your API key at openrouter.ai/keys

Other Models from OpenAI

Compare performance with other models from the same creator

ModelLatencyCost/1MScore
GPT-5.54.9s$17.50100%
GPT-5.5 Pro28.9s$105.0077%
GPT-5.220.7s$7.8869%
GPT-5.3 Chat2.5s$7.8865%
GPT-5.4 Mini1.2s$2.6357%
GPT-5 Mini24.2s$1.1357%
GPT-5.41.8s$8.7549%
gpt-oss-20b13.6s$0.0941%
GPT-5 Nano33.8s$0.2230%
GPT-5.4 Nano—$0.73—
GPT-5.4 Pro—$105.00—
GPT-5.3-Codex2.7s$7.88—
GPT Audio—$6.25—
GPT Audio Mini—$1.50—
GPT-5.2-Codex21.7s$7.88—
GPT-5.2 Chat9.5s$7.88—
GPT-5.2 Pro47.7s$94.50—
GPT-5.1-Codex-Max25.2s$5.63—
GPT-5.130.5s$5.63—
GPT-5.1 Chat5.5s$5.63—
GPT-5.1-Codex16.6s$5.63—
GPT-5.1-Codex-Mini12.6s$1.13—
gpt-oss-safeguard-20b2.4s$0.19—
GPT-5 Image Mini31.8s$2.25—
GPT-5 Image49.2s$10.00—
o3 Deep Research362.7s$25.00—
o4 Mini Deep Research151.0s$5.00—
GPT-5 Pro332.1s$67.50—
GPT-5 Codex19.7s$5.63—
GPT-4o Audio—$6.25—
GPT-5 Chat5.9s$5.63—
GPT-558.6s$5.63—
o3 Pro122.5s$50.00—
o4 Mini High26.8s$2.75—
o319.0s$5.00—
o4 Mini18.2s$2.75—
GPT-4.120.7s$5.00—
GPT-4.1 Mini11.5s$1.00—
GPT-4.1 Nano4.5s$0.25—
o1-pro95.3s$375.00—
o3 Mini High13.8s$2.75—
o3 Mini16.0s$2.75—
o123.4s$37.50—
GPT-4o (2024-11-20)12.4s$6.25—
GPT-4o (2024-08-06)9.5s$6.25—
GPT-4o5.8s$12.00—
GPT-4o (2024-05-13)4.7s$10.00—
GPT-4o-mini Search Preview6.5s$0.38—
GPT-4o Search Preview11.2s$6.25—
GPT-4 Turbo26.5s$20.00—
GPT-4 Turbo Preview15.8s$20.00—
GPT-3.5 Turbo 16k4.0s$3.50—
GPT-412.1s$45.00—
Codex Mini—$3.75—
ChatGPT-4o7.4s$10.00—
GPT-3.5 Turbo (older v0613)13.7s$1.50—
GPT-4 Turbo (older v1106)18.1s$20.00—
GPT-3.5 Turbo Instruct4.4s$1.75—
GPT-4 (older v0314)16.8s$45.00—
GPT-3.5 Turbo3.8s$1.00—
GPT Chat Latest—$17.50—
GPT-5.4 Image 241.3s$11.50—
GPT-4o-mini11.0s$0.38—
GPT-4o-mini (2024-07-18)13.2s$0.38—