What are the strengths of this model?

Advanced coding capabilities Long 400K context understanding Latest design from OpenAI

What are the weaknesses of this model?

Is closed-source Usage subject to license restrictions Limited publicly available information

What are the best use cases?

Automatic generation of complex code Analysis of large repositories Advanced program modification

Back to Models

OpenAIProprietary

OpenAI GPT-5.1 Codex

Name: OpenAI GPT-5.1 Codex
Author: OpenAI

OpenAI GPT-5.1 Codex is a programming-specialized foundation model developed by OpenAI. It features an extensive 400K context window.

Parameters

Undisclosed

Context Window

400K

License

Proprietary

Release Date

2025-11-14

API Pricing

API pricing for this model is not yet available

Strengths

・Advanced coding capabilities
・Long 400K context understanding
・Latest design from OpenAI

Weaknesses

・Is closed-source
・Usage subject to license restrictions
・Limited publicly available information

Use Cases

・Automatic generation of complex code
・Analysis of large repositories
・Advanced program modification

Deep Analysis

Release Date

November 19, 2025

Context Window

400K tokens

Max Output

128K tokens

Input Price

$1.25 / 1M tokens

Output Price

$10.00 / 1M tokens

Cache Read

$0.125 / 1M tokens

SWE-bench Verified

73.7%

SWE-Lancer

66.3%

Terminal-Bench 2.0

52.8%

Knowledge Cutoff

September 2024

Strengths

・Strong coding benchmarks: SWE-bench Verified 73.7%, SWE-Lancer 66.3%, Terminal-Bench 52.8%
・Large 400K context window for processing entire codebases
・Supports tool calling, reasoning, structured output, and attachments
・Available across 20+ providers with competitive pricing
・Good agent score (100/100 on modelcompare.dev) and production-readiness (96/100)

Weaknesses

・Superseded by GPT-5.1-Codex-Max (SWE-bench 77.9%) and GPT-5.2-Codex
・No temperature control — limited fine-tuning of output randomness
・Proprietary/closed-source — only OpenAI and approved partners can serve it
・Higher cost than GPT-5.1 Codex Mini ($1.25/$10 vs $0.25/$2)
・Knowledge cutoff of September 2024 is older than newer models

Competitor Comparison

Model	Arena	SWE	GPQA	Price
Claude Sonnet 4.5	~1420	~72%	~89%	$3/$15 per 1M tokens
GPT-5.1-Codex-Max	~1410	77.9%	~88%	$1.25/$10 per 1M tokens
Gemini 3 Pro	~1420	76.2%	~90%	$3.50/$10.50 per 1M tokens
GPT-5-Codex	~1390	~70%	~85%	$1.25/$10 per 1M tokens

Overview

GPT-5.1 Codex is OpenAI's coding-specialized model released November 19, 2025, achieving SWE-bench Verified 73.7% and Terminal-Bench 2.0 52.8% with a 400K context window. It replaced GPT-5-Codex as the default coding model before being itself superseded by GPT-5.1-Codex-Max just days later. At $1.25/$10 per 1M tokens, it remains available as a mid-tier coding option.

Benchmarks & Performance

GPT-5.1 Codex achieves SWE-bench Verified 73.7%, SWE-Lancer IC SWE 66.3%, and Terminal-Bench 2.0 52.8%. On modelcompare.dev, it scores 100/100 for agents, 96/100 for production-readiness, 94/100 for vision, and 92/100 for coding. Cost efficiency scenarios: RAG answer (1K) costs $11.25 ($0.01 each), code review (1K PRs) costs $20.00 ($0.02 each). The model supports tool calling, reasoning, and structured output but lacks temperature control.

Detailed Comparison

GPT-5.1 Codex trails GPT-5.1-Codex-Max (77.9%) and Claude Opus 4.5 (80.9%) on SWE-bench but offers competitive pricing. It beats Gemini 3 Pro (76.2%) on SWE-bench at a lower price ($1.25/$10 vs $3.50/$10.50). Compared to GPT-5-Codex, it offers improved benchmarks at the same price. Claude Sonnet 4.5 ($3/$15) is more expensive with a similar SWE-bench score. The model is available across 20+ providers, with some offering discounts (OpenCode Zen at $1.07/$8.50).

Community Feedback

GPT-5.1 Codex was well-received as a strong coding model but was quickly overshadowed by Codex-Max. Developers appreciate the 400K context window and agent capabilities. The lack of temperature control is a common complaint for creative coding tasks. It remains popular for mid-tier coding workloads where Codex-Max's additional cost isn't justified. The model is widely available across providers, making it accessible for diverse deployment scenarios.

Use Cases

Best for: code review and refactoring, software engineering tasks, structured code generation, agentic coding workflows, multi-file project analysis, and front-end/interactive development. The 400K context window is ideal for processing entire repositories. For simpler or higher-volume tasks, GPT-5.1 Codex Mini ($0.25/$2) is more cost-effective. For complex autonomous coding, GPT-5.1-Codex-Max offers better accuracy. For interactive pair programming, Claude Code may be preferred due to lower code churn.

Latest News

Released November 19, 2025, replacing GPT-5-Codex. Was itself replaced as default by GPT-5.1-Codex-Max on the same day. GPT-5.2-Codex and GPT-5.3-Codex have since been released as newer coding models. Available on OpenAI API and 20+ third-party providers.

Sources

Analysis generated: 2026-05-24