Skip to content
llmwaves

gpt-oss-120b

OpenAI · released Aug 5, 2025

Open weightsReasoningTool useStructured output

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...

Compare this modelSource listing

Specification

Context window
131K
Max output
131K
Knowledge cutoff
Jun 30, 2024
Parameters
116.8B
Licence
apache-2.0
Serving providers
17
Moderated
No
Uptime
100.0%

Intelligence

24.1

66th percentile

Coding

30.4

Coding Index

Agentic

13.4

Agentic Index

Output speed

773 t/s

Median across providers

Latency

205ms

Time to first token

Cost per task

$0.02

Estimated

Benchmarks

Where the score comes from

The Intelligence Index is a composite. These are the underlying evaluations this model was actually measured on.

Evaluation scores

Percentage correct · higher is better

  • AIME 2025
    93.4%
  • LiveCodeBench
    87.8%
  • MMLU-Pro
    80.8%
  • GPQA Diamond
    78.2%
  • IFBench
    69.0%
  • τ²-bench (Telecom)
    65.8%
  • AA-LCR (long context)
    51.0%
  • SciCode
    38.9%
  • Terminal-Bench Hard
    23.5%
  • Humanity's Last Exam
    19.6%

An evaluation missing from this list was not run for this model — it is not a zero.

View as table
EvaluationScore
AIME 202593.4%
LiveCodeBench87.8%
MMLU-Pro80.8%
GPQA Diamond78.2%
IFBench69.0%
τ²-bench (Telecom)65.8%
AA-LCR (long context)51.0%
SciCode38.9%
Terminal-Bench Hard23.5%
Humanity's Last Exam19.6%

Against its peers

Intelligence Index · this model highlighted, nearest peers in grey

Peers are the models sitting closest on the Intelligence Index — the set you would realistically choose between.

View as table
ModelIntelligence
Qwen3 Max24.5
Ling 3.0 Tiny (free)24.3
gpt-oss-120b24.1
Claude Haiku 4.524.1
Kimi K2 090524.0
o123.9
GLM 4.623.4
GLM 4.7 Flash23.3

Percentile among all indexed models

Intelligence66th
Coding42th
Agentic26th
Mathematics95th
Terminal work63th
Arena Elo16th

Pricing

What it costs to run

List prices per million tokens, plus what one representative task works out to.

List price

Input / 1M tokens
$0.037
Output / 1M tokens
$0.17
Cached input / 1M
Not offered
Blended 3:1
$0.07

One task, estimated

$0.02

Input tokens
50,000
Output tokens
80,000
Profile
Reasoning

Estimated from list pricing: 50K input tokens plus 80K output tokens for reasoning models (25K for non-reasoning).

Serving providers

Speed and latency figures are medians across these providers, so a widely-served model reports a blend rather than any single endpoint.

Arena

Head-to-head generation quality

Elo from pairwise judgements, broken out by the kind of thing the model was asked to build.

Overall Elo
1,034
Win rate
40.6%
Strongest at
Game development
Tournaments
443

Elo by category

Dot position on a 950–1,050 scale · Elo has no meaningful zero

Agent categories (full-stack apps, mobile apps) are only scored for models tested in agent harnesses.

View as table
CategoryEloWin rate
Game development103440.6%
Data visualisation102745.1%
Websites99032.5%
UI components95935.5%
3D scenes95629.4%

Same creator

More from OpenAI

GPT-5.6 Sol

Intelligence
60.9
Speed
46 t/s
Cost per task
$2.65

GPT-5.6 Terra

Intelligence
56.6
Speed
56 t/s
Cost per task
$0.53

GPT-5.5

Intelligence
56.3
Speed
85 t/s
Cost per task
$2.65

GPT-5.4

Intelligence
53.1
Speed
120 t/s
Cost per task
$1.32

GPT-5.6 Luna

Intelligence
52.3
Speed
88 t/s
Cost per task
$0.05

GPT-5.3-Codex

Intelligence
45.5
Speed
61 t/s
Cost per task
$1.21