gpt-⁠oss-120b.

Among the lowest scores.

OpenAI · Open weights · Released 5 August 2025
Data from 13 November 2026.

  • Score56th of 58

    4

    Artificial Analysis's Intelligence Index. The median is 27.

  • Cost14th of 51

    $0.35

    What each task of that index costs. The median is $1.03.

  • Time14th of 50

    7.9 min

    How long each task of that index takes. The median is 14.6 min.

  • Context51st of 55

    131 thousand

    Tokens it reads at once, as 3 others do.

In short

It scores 4 on Artificial Analysis's Intelligence Index, 56th of the 58 models it scores. The median is 27.

Each task of that index costs it $0.35 and takes it 7.9 minutes on average. The median of the 58 is $1.03 and 14.6 minutes.

OpenAI does not sell it per token, and it reads up to 131 thousand tokens at once.

Its weights are published, so it can run on your own servers.

Its data

Maker
OpenAI
Weights
Open
Effort measured
High
Price per million tokens
No data
Context
131 thousand tokens
Released
5 Aug 2025

The score, the cost and the time per task and its tests from Artificial Analysis, read on 13 November 2026 at its highest effort. The price, the context and the release date from OpenAI's page. The table of the 58.

Its tests

Coding

  • FrontierCode 2.0 Changes to real open-⁠source projects. 51st of 58.0.0%
  • Terminal-⁠Bench 4.0 Complex work in the terminal, on its own. 48th of 58.0.0%

Reasoning

  • HLE Diamond Expert-⁠level questions. 56th of 58.3.8%
  • CritPt v2, hard subset Physics research problems. 46th of 58.4.0%

Questions.

  • How much does gpt-⁠oss-120b cost?

    OpenAI does not sell it per token.

  • How much text does it read at once?

    Up to 131 thousand tokens, according to OpenAI.

  • Can it run on your own servers?

    Yes. Its weights are published, so it can be downloaded and run on your own servers.

  • Which one for your company's software?

    It depends on the job. We cover it in “Which AI Model to Use to Build Your Company's Software”.

Let's talk.

Tell us what you need and one of the partners who write the code gets back to you.