Side-by-side comparison · models
GPT-4o (2024-08-06) vs GPT-5 Codex
What we have checked about each, line by line. Every value shows where it came from and when we last confirmed it.
Models
GPT-4o (2024-08-06)
The 2024-08-06 version of GPT-4o offers improved performance in structured outputs, with the ability to supply a JSON schema in the respone_format. Read more [here](https://openai.
Models
GPT-5 Codex
GPT-5-Codex is a specialized version of GPT-5 optimized for software engineering and coding workflows.
Checked facts
Line by line
| Measure | GPT-4o (2024-08-06) | GPT-5 Codex |
|---|---|---|
| Input price | $2.50 per million tokens | Not checked yet |
| Output price | $10.00 per million tokens | Not checked yet |
| Price on azure | $1.25 per million cached input tokens | Not checked yet |
| Price on azure | $2.50 per million input tokens | Not checked yet |
| Price on azure | $10.00 per million output tokens | Not checked yet |
| Price on openai | $1.25 per million cached input tokens | Not checked yet |
| Price on openai | $2.50 per million input tokens | Not checked yet |
| Price on openai | $10.00 per million output tokens | Not checked yet |
| Latest version | openai/gpt-4o-2024-08-06 | Not checked yet |
| Time to first token (ms) | 1,990 | 8,380 |
| Time to first token (ms) on azure | 726 | Not checked yet |
| Time to first token (ms) on openai | 565 | Not checked yet |
| Output speed (tokens a second) | 119.6 | 165.7 |
| Output speed (tokens a second) on azure | 61 | Not checked yet |
| Output speed (tokens a second) on openai | 39 | Not checked yet |
| Tokens served on OpenRouter in a day | 376.5M (for 10 Nov 2026) | Not checked yet |
| Tool calls on OpenRouter in a day | 702 (for 10 Nov 2026) | Not checked yet |
| Artificial Analysis Intelligence Index | 7.7 | 24.9 |
| LMArena text rating (Elo) | 1,300 | Not checked yet |
| LMArena vision rating (Elo) | 1,137 | Not checked yet |
| Gpqa diamond | 49.2% (for 6 Aug 2024) | Not checked yet |
| Frontiermath tiers 1 3 v2 private | 0.4% (for 6 Aug 2024) | Not checked yet |
| Simpleqa verified | 26% (for 6 Aug 2024) | Not checked yet |
| Aider polyglot | 23.1% (for 6 Aug 2024) | Not checked yet |
| Cadeval | 26% (for 6 Aug 2024) | Not checked yet |
| Chess puzzles | 6% (for 6 Aug 2024) | Not checked yet |
| Frontiermath 2025 02 28 private | 0.3% (for 6 Aug 2024) | Not checked yet |
| Math level 5 | 53.3% (for 6 Aug 2024) | Not checked yet |
| Mmlu | 84.3% (for 6 Aug 2024) | Not checked yet |
| Simplebench | 17.8% (for 6 Aug 2024) | Not checked yet |
| Otis mock aime 2024 2025 | 6.4% (saturated benchmark: the top models all score near the ceiling) (for 6 Aug 2024) | Not checked yet |
| Terminal bench | Not checked yet | 44.3% (for 15 Sept 2025) |
| Weirdml | Not checked yet | 54.5% (for 15 Sept 2025) |
A line appears only when at least one record has a value we confirmed from a source. Values we have not confirmed are left out, never guessed. Records of different types share fewer lines.
Add a third
Compare with one more record
How this works
Share it, or build your own
This page works for any two or three records in the directory, of any type. The link stays the same whichever order you name them in, so you can share it as it is. Change the records in the address, for example /directories/vs/claude-code-vs-cursor, or start from our editorial comparisons.