LLM Reference

Codex 1

Released
2025-05-16
Last refreshed
2026-04-18
Status
Researched 187d ago
DeprecatedProprietaryCommercial use: conditionalCodingAgentsLong context

Codex 1 is a legacy integration reference; keep it only while you identify a current replacement.

Use it for

  • Teams maintaining an existing integration
  • Workloads that can use a 192k context window

Do not use it for

  • New production launches
  • Cost-sensitive launches that need sourced token pricing
  • Vision or document-understanding workloads
Specifications
Released
2025-05-16
Context
192k
Architecture
Decoder Only
Specialization
code
Openness
Proprietary
License
ProprietaryCommercial use: conditional
Weights
Not released
Code
Unknown
Training
Reinforcement Learning
Created by
OpenAIOpenAI

Cutting-edge research and development.

San Francisco, California, United States
Founded 2015
Website
Pricing

No tracked provider token pricing is available yet.

Links
Website

About

OpenAI's primary Codex model for software engineering, built on o3 architecture. Trained via reinforcement learning on real-world coding tasks. Features sandboxed container execution, GitHub PR export, diff review, and multi-file reasoning. Default model for the OpenAI Codex cloud coding agent, available on Plus, Pro, Business, and Enterprise tiers.

Codex 1 is a proprietary model in the OpenAI Codex family. The structured metadata tracks a 192k-token context window, reasoning, and code execution. Headline tracked benchmarks include SWE-bench Verified 72.1.

Top use-case fit: coding, agents, and build tasks

Coding

1 relevant benchmark in the decision map.

Agents

1 relevant benchmark in the decision map.

Long context

Included by capability and metadata signals in the decision map.

Provider price ladder

No tracked provider token pricing is available for this model yet.

Capabilities

ReasoningCode Execution

Benchmark peer barsfor Coding

SWE-bench VerifiedRank 55 of 80
Claude Fable 5
96.0
Claude Mythos Preview
93.9
Claude Opus 4.8
88.6
Claude Opus 4.7
87.6
Codex 1current
72.1

Benchmark scores(1)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
BenchmarkScoreVersionSource
SWE-bench Verified72.12025-05https://openai.com/index/introducing-codex/

Migration checks

No linked migration route is available for this model yet.

Compare Codex 1 with other models

Show all 32 popular comparisonssorted by 7-day search impressions
Codex 1 vs Llama 3.1 Nemotron Nano VL 8B v114Codex 1 vs text-davinci12Codex 1 vs Llama 3.1 Swallow 8B Instruct12Codex 1 vs Teuken 7B Instruct11Codex 1 vs Together AI - Llama 3 8B Lite11Codex 1 vs Mistral Large 3 675B Instruct10Codex 1 vs Seed-OSS 36B Instruct10Codex 1 vs ELYZA Japanese Llama 2 7B9Codex 1 vs Sarvam-M Multilingual Hybrid9Codex 1 vs Phi-4 Mini Flash Reasoning9Codex 1 vs Together AI - Gemma 3n-e4B8Codex 1 vs Falcon 180B8Codex 1 vs Qwen2-7B-Instruct7Codex 1 vs Llama 3 Swallow 70B Instruct7Codex 1 vs Llama 3.2 NV RerankQA 1B v26Codex 1 vs Mistral Small 34Codex 1 vs Colosseum 355B Instruct4Codex 1 vs Mistral 7B v0.14Codex 1 vs Llama 3.1 Nemotron Nano 4B v1.13Codex 1 vs ShieldGemma 9B3Codex 1 vs Llama 2 7B Chat3Codex 1 vs Aquila 2 7B3Codex 1 vs GPT-2 Medium2Codex 1 vs Nemotron Mini Hindi 4B Instruct2Codex 1 vs Marin 8B Instruct2Codex 1 vs Granite Guardian 3.0 8B1Codex 1 vs Llama Guard 2 8B1Codex 1 vs Bielik 11B v2.6 Instruct1Codex 1 vs Falcon 3 7B Instruct1Codex 1 vs Gemma 2 9B SahabatAI Instruct1Codex 1 vs Stockmark 2 100B Instruct1Codex 1 vs Magistral Small 25061

Frequently asked questions

What is the context window of Codex 1?

Codex 1 has a context window of 192k tokens.

When was Codex 1 released?

Codex 1 was released on 2025-05-16.

What benchmarks has Codex 1 been tested on?

Codex 1 has been evaluated on 1 benchmark, including SWE-bench Verified.