GPT-5.3-Codex

Frontier

OpenAI•Released on 2025-12-18

OpenAI's coding-optimized model, surpassing Claude on SWE-bench. HN users praise its coding value and much more generous quotas than Claude. Ideal for intensive coding work.

71
Overall Score

Core Specs

256K
Context Window
64K
Max Output
✓
Reasoning
✗
Open Source
Multimodal Support
text

Scenario Scores

Coding
#4 / 1493
Writing
#14 / 1460
Translation
#14 / 1465
Data Analysis
#12 / 1475
Conversation
#14 / 1460
Image Gen
—

User Feedback Highlights

Based on community feedback. Hover to see original reviews.

+ SWE-bench leader− Text-only+ Coding-optimized− Weak at creative tasks+ Generous quotas+ Great value− Requires Codex-specific API
Sentiment:👍 61%😐 28%👎 11%

Pros & Cons

Pros

  • +Coding-optimized
  • +Great value
  • +Generous quotas
  • +SWE-bench leader

Cons

  • −Text-only
  • −Weak at creative tasks
  • −Requires Codex-specific API

Reliability

SLA99.9%
Incidents (30d)3
与 GPT-5.4 共享基础设施
View Status Page →

Pricing

Input (per 1M tokens)$1.75
Output (per 1M tokens)$14.00
Subscription$20/month
Free trial available
Updated on 2026-03-06

Tools Supporting This Model

Codex CLI
OpenAI

Compare with Others

GPT-5.3-Codex vs GPT-5.4 →GPT-5.3-Codex vs GPT-5.4 Thinking →GPT-5.3-Codex vs Claude Opus 4.6 →

Benchmarks

sweBench76.3%
humanEval97.2%
aiderPolyglot88%