Skip to content

OpenAI Proprietary Released Feb 5, 2026 Updated

GPT-5.3 Codex

ToolsGuide Scores, rankings, test results and price.

Model summary

  • Coding #86 / 147

    293 / 1,000

    ToolsGuide Score· votes 263, tests 341

  • Price per 1M tokens

    In $1.75· Out $14.00

    $1.75

    Input · median coding model $0.75

Summary

GPT-5.3 Codex is OpenAI's model, released Feb 5, 2026. It's ranked in 1 category and does best at coding: #86 of 147, with a ToolsGuide Score of 293 against a median of 316.

It costs $1.75 per million input tokens and $14.00 per million output tokens, about 2.3× the median coding model ($0.75).

It has published results on 6 tests and has the top score on 1 of them, including Cybersecurity CTFs.

Specifications

Creator
OpenAI
Licence
Proprietary
Released
Feb 5, 2026
Context window
400K tokens
Input price
$1.75 / 1M
Output price
$14.00 / 1M
Tests reported
6
Settings ranked
GPT-5.3-Codex (Codex Harness)

Model comparison

Against the ten best coding models.

Coding score

ToolsGuide Score out of 1,000 · Higher is better

Input price

USD per 1M input tokens · Lower is better

Leaderboards

Where GPT-5.3 Codex sits among the top 20 in each category.

Coding

#86 of 147 · ToolsGuide Score out of 1,000, where 1,000 is a flawless model

Full coding leaderboard →

By task

Scored from votes on each kind of prompt alone, on the same 1,000-point scale.

Score against price

Coding · the 10 priced models ranked closest · up and to the left is better value

Coding

GPT-5.3-Codex (Codex Harness) · hover a bar for its rank

Test results

Every test GPT-5.3 Codex has a published result on, out of 100%, with its rank among the models that took it.

  • SWE-Bench Pro

    #33 / 61

    Harder, longer software engineering tasks from real codebases.

    56.8%

    Self-reported · In the coding score

  • Terminal-Bench 2.0

    #4 / 54

    Getting real work done in a command-line terminal.

    77.3%

    Self-reported

  • LiveBench

    #23 / 38

    LiveBench is a challenging, contamination-limited LLM benchmark that addresses test set contamination by releasing new questions monthly based on recently-released datasets, arXiv papers, news articles, and IMDb movie synopses. It comprises tasks across math, coding, reasoning, language, instruction following, and data analysis with verifiable, objective ground-truth answers.

    72.8%

    Self-reported · In the text score

  • OSWorld-Verified

    #22 / 28

    OSWorld-Verified is a verified subset of OSWorld, a scalable real computer environment for multimodal agents supporting task setup, execution-based evaluation, and interactive learning across Ubuntu, Windows, and macOS.

    64.7%

    Self-reported

  • SWE-Lancer (IC-Diamond subset)

    #2 / 6

    SWE-Lancer (IC-Diamond subset) is a benchmark of real-world freelance software engineering tasks from Upwork, ranging from $50 bug fixes to $32,000 feature implementations. It evaluates AI models on independent engineering tasks using end-to-end tests triple-verified by experienced software engineers, and includes managerial tasks where models choose between technical implementation proposals.

    81.4%

    Self-reported

  • Cybersecurity CTFs

    #1 / 3

    Cybersecurity Capture the Flag (CTF) benchmark for evaluating LLMs in offensive security challenges. Contains diverse cybersecurity tasks including cryptography, web exploitation, binary analysis, and forensics to assess AI capabilities in cybersecurity problem-solving.

    77.6%

    Self-reported