GPT-5.3-Codex
OpenAI released 2026-02-24 proprietary
No independent quality score.
No benchmark we track has measured this model. That is not the same as measuring it and finding it wanting — we simply cannot rank it, so we do not.
Our take
editorial — not a measurementA coding-specialised variant at $1.75/$14 with a 400k window. The asymmetry is the story: output costs 8x input, which is unusual and punishing for code generation specifically, since codegen is an output-heavy workload. Compare carefully against Kimi K2.7 Code ($0.67/$3.40) and Claude Sonnet 5 ($2/$10) on cost-per-completed-task rather than per token. Worth it if you are already inside OpenAI's Codex tooling; hard to justify on price alone.
Strengths
- Purpose-tuned for coding workflows
- 400k context window
- Cached input at $0.175 makes repeated-repo context cheap
Weaknesses
- $14 output on an output-heavy workload type
- 8:1 output-to-input ratio is worse than most peers
- Undercut heavily by Kimi K2.7 Code on price
Reach for it when
- Teams already standardised on OpenAI Codex tooling
- Repo-scale context with heavy cache reuse
Avoid it if
- You generate large volumes of code output
- You are price-comparing against open-weight coding models
Sources: developers.openai.com
Every price dimension
| Input | $1.75/M |
|---|---|
| Output | $14.00/M |
| Cached input | $0.175/M |
| Web search | $0.01/call |
| Batch discount | 50% |
Independent scores
No independent benchmark has measured this model.
Best rankings by task
- godotgamedev #24of 43 1124
- asciiart #31of 83 1178
- androidnative #31of 54 1084
Capability
| Context window | 400K |
|---|---|
| Max output | 128K |
| Input modes | text, image, file |
| Tool use | yes |
| Reasoning | optional |
| Open weights | no |
Provenance
| Price source | openrouter.ai |
|---|---|
| Fetched | 2026-08-24 |
| Quality data | unrated |
| Cross-checked | vendor page |
GPT-5.3-Codex is OpenAI’s most advanced agentic coding model, combining the frontier software engineering performance of GPT-5.2-Codex with the broader reasoning and professional knowledge capabilities of GPT-5.2. It achieves state-of-the-art results...