Updated March 26, 2026· Based on independent benchmark data
GPT-5.2 Codex (xhigh) and GPT-5.4 mini (xhigh) are virtually tied on intelligence (49.0 vs 48.1). GPT-5.4 mini (xhigh) is 2.3x cheaper at $0.75/1M tokens vs $1.75/1M. For speed, GPT-5.4 mini (xhigh) wins at 218 tok/s vs 99 tok/s.
| Metric | GPT-5.2 Codex (xhigh) | GPT-5.4 mini (xhigh) |
|---|---|---|
| Intelligence Score | 49.0 | 48.1 |
| Coding Score | 43.0 | 51.5 |
| Math Score | N/A | N/A |
| Speed (tok/s) | 99 tok/s | 218 tok/s |
| Latency (TTFT) | 6.86s | 7.45s |
| Input Price / 1M tokens | $1.75 | $0.75 |
| Output Price / 1M tokens | $14 | $4.50 |
| Context Window | N/A |
GPT-5.2 Codex (xhigh) and GPT-5.4 mini (xhigh) perform similarly on overall intelligence, scoring 49.0 and 48.1 respectively. For coding tasks, GPT-5.4 mini (xhigh) has the edge with a coding score of 51.5 vs 43.0.
GPT-5.4 mini (xhigh) generates output significantly faster at 218 tok/s compared to GPT-5.2 Codex (xhigh)'s 99 tok/s, making it 2.2x faster for streaming responses. Time to first token is 6.86s for GPT-5.2 Codex (xhigh) vs 7.45s for GPT-5.4 mini (xhigh), which affects perceived responsiveness in interactive applications.
GPT-5.4 mini (xhigh) is more affordable at $0.75/1M input tokens ($4.50/1M output), while GPT-5.2 Codex (xhigh) costs $1.75/1M input ($14/1M output). That makes GPT-5.2 Codex (xhigh) 2.3x more expensive per token, which can add up significantly at scale. For a typical workload of 100 requests per day at 2,000 tokens each, GPT-5.2 Codex (xhigh) would cost approximately $10.50/month vs $4.50/month for GPT-5.4 mini (xhigh) in input costs alone.
Choose GPT-5.4 mini (xhigh) when you need stronger coding performance (51.5), faster output (218 tok/s), lower cost.
GPT-5.4 mini (xhigh) scores higher on coding benchmarks (51.5 vs 43.0), making it the better choice for programming tasks.
GPT-5.4 mini (xhigh) is cheaper at $0.75/1M input tokens vs $1.75/1M for GPT-5.2 Codex (xhigh).
GPT-5.4 mini (xhigh) is faster, producing output at 218 tok/s compared to GPT-5.2 Codex (xhigh)'s 99 tok/s.
No, GPT-5.2 Codex (xhigh) does not support image input. Neither model supports image input.
Data last synced: March 26, 2026
| N/A |
| Max Output Tokens | N/A | N/A |
| Input Modalities | Text | Text |
| Output Modalities | Text | Text |
| Free Tier | No | No |
Both models perform similarly on intelligence benchmarks. Choose based on specific needs: pricing, speed, context window, or provider ecosystem.