# GPT-5.3-Codex

135 Hacker News commenters recommended GPT-5.3-Codex and 78 criticized it, November 2025 to November 2026. It was recommended in 11 of 13 months; owners who said how long report a median of 0.17 years (4 owners).

**Verdict:** A careful OpenAI coding model praised for long tasks, orchestration, verification and code review at good value; critics report silent routing to weaker models and robotic, hard-to-follow explanations.

## Why people recommend it

- Excellent overall coding ability (8 comments): "Since the performance of the latest models (Opus 4.6 max-effort, gpt-5.3-Codex) is nothing short of astonishing." ([source](https://news.ycombinator.com/item?id=47015110)) "This is literally stuff of wonders and magic that redefines how we interface with computers and code." ([source](https://news.ycombinator.com/item?id=46904469)) "GPT 5.5 codex was shockingly good. Achieved 90% perfect on first try in about a fourth of the time." ([source](https://news.ycombinator.com/item?id=47896320))
- Handles long tasks and orchestration (4 comments): "Gpt-5.3-codex is miles better than 5.4 in that regard. It's better at orchestration, and does the things that it said it did." ([source](https://news.ycombinator.com/item?id=47881791)) "the Opus 4.5/4.6 and GPT-5.2/5.3 models really have represented a step-change in how good they are at running long tasks" ([source](https://news.ycombinator.com/item?id=46993314)) "gave me the confidence to load in a ridiculous number of tasks every day and let it rip" ([source](https://news.ycombinator.com/item?id=47082391))
- Careful, verifies and reviews code (5 comments): "Somehow it's just way more careful than the others, and also much better at empirical verification of its hypothesis, writing tests, etc." ([source](https://news.ycombinator.com/item?id=48571458)) "when GPT-5.3 Codex looked at it, it said that speaker A and speaker B were switched in the LoRA code" ([source](https://news.ycombinator.com/item?id=47260601)) "surprisingly adept at performing a high level feature branch review and catching issues" ([source](https://news.ycombinator.com/item?id=47472475))
- Good value versus competitors (3 comments): "subjectively think it does better job than Opus 4.6/4.7, with a fraction of the cost" ([source](https://news.ycombinator.com/item?id=48319921)) "its pricing is really good in production (due to how their token caching works)" ([source](https://news.ycombinator.com/item?id=48313290)) "which has more than double context size, seems just as smart and costs the same" ([source](https://news.ycombinator.com/item?id=47055868))

## What people warn about

- Reliability degradation over time (5 comments): "I have noticed this degradation of 5.5 reliability to what, in my experience, I consider Claude-level of reliability since early June." ([source](https://news.ycombinator.com/item?id=48791064)) "GPT-5.5 Codex reasoning-token clustering at 516/1034/1552 may be leading to degraded performance on complex tasks" ([source](https://news.ycombinator.com/item?id=48849569)) "occasionally it will seemingly short circuit and think for exactly 516 tokens, and return the wrong result" ([source](https://news.ycombinator.com/item?id=48789957))
- Silently routed to weaker model (2 comments): "Pulling a switcheroo on the user behind the scenes, whatever the justification, is another issue" ([source](https://news.ycombinator.com/item?id=46970395)) "getting their requests silently routed to a less capable model without telling the user at all" ([source](https://news.ycombinator.com/item?id=46971951))
- Output often needs iteration (3 comments): "Some were definitely below par, and I recall having to iterate on the generated code more than I wanted to." ([source](https://news.ycombinator.com/item?id=48795647)) "I've run multiple gpt-5 codex concurrent sessions in the cloud, but I didn't accept one thing they did." ([source](https://news.ycombinator.com/item?id=45987056)) "I'm not seeing this translating to useful code. It failed with the first few requests I gave it." ([source](https://news.ycombinator.com/item?id=46984470))
- Claims done without clarity (2 comments): "When I've used codex (it has been getting better) but back in the day it would just do things and say it's done and you're just sitting there wondering "wtf are you doing?"" ([source](https://news.ycombinator.com/item?id=46907057)) "5.3 is good but talks like a robot, it’s too hard to understand what exactly it’s talking about." ([source](https://news.ycombinator.com/item?id=48330315))

## Alternatives hackers recommend

- [Cursor Composer](https://hackerapproved.com/p/composer-2-5.md): 214 recommend
- [Qwen3-Coder](https://hackerapproved.com/p/qwen3-coder-next.md): 165 recommend
- [Meta Muse Glimmer 30B](https://hackerapproved.com/p/meta-muse-glimmer.md): 70 recommend
- [GPT-5.2-Codex](https://hackerapproved.com/p/gpt-5-2-codex.md): 54 recommend
- [Laguna S 2.1](https://hackerapproved.com/p/laguna-s-2-1.md): 43 recommend

## Where to get it

- [Visit the official site](https://openai.com/codex/)

## Cite as

Hacker Approved, "GPT-5.3-Codex", https://hackerapproved.com/p/gpt-5-3-codex/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).