Updated March 26, 2026· Based on independent benchmark data
GPT-5.4 (xhigh) leads in intelligence with a score of 57.2 vs 51.3.
| Metric | GPT-5.2 (xhigh) | GPT-5.4 (xhigh) |
|---|---|---|
| Intelligence Score | 51.3 | 57.2 |
| Coding Score | 48.7 | 57.3 |
| Math Score | 99.0 | N/A |
| Speed (tok/s) | 70 tok/s | 77 tok/s |
| Latency (TTFT) | 75.24s | 181.82s |
| Input Price / 1M tokens | $1.75 | $2.50 |
| Output Price / 1M tokens | $14 | $15 |
| Context Window | N/A | N/A |
| Max Output Tokens | N/A | N/A |
| Input Modalities | Text | Text |
GPT-5.4 (xhigh) outperforms GPT-5.2 (xhigh) on the intelligence index with a score of 57.2 compared to 51.3. For coding tasks, GPT-5.4 (xhigh) has the edge with a coding score of 57.3 vs 48.7.
Both models deliver similar output speeds: GPT-5.2 (xhigh) at 70 tok/s and GPT-5.4 (xhigh) at 77 tok/s. Time to first token is 75.24s for GPT-5.2 (xhigh) vs 181.82s for GPT-5.4 (xhigh), which affects perceived responsiveness in interactive applications.
GPT-5.2 (xhigh) is more affordable at $1.75/1M input tokens ($14/1M output), while GPT-5.4 (xhigh) costs $2.50/1M input ($15/1M output). For a typical workload of 100 requests per day at 2,000 tokens each, GPT-5.2 (xhigh) would cost approximately $10.50/month vs $15.00/month for GPT-5.4 (xhigh) in input costs alone.
Choose GPT-5.4 (xhigh) when you need higher intelligence (57.2), stronger coding performance (57.3).
GPT-5.4 (xhigh) scores higher on coding benchmarks (57.3 vs 48.7), making it the better choice for programming tasks.
GPT-5.2 (xhigh) is cheaper at $1.75/1M input tokens vs $2.50/1M for GPT-5.4 (xhigh).
GPT-5.4 (xhigh) is faster, producing output at 77 tok/s compared to GPT-5.2 (xhigh)'s 70 tok/s.
No, GPT-5.2 (xhigh) does not support image input. Neither model supports image input.
Data last synced: March 26, 2026
| Output Modalities |
| Text |
| Text |
| Free Tier | No | No |
It depends on your priorities. GPT-5.4 (xhigh) scores higher on intelligence (57.2), but GPT-5.2 (xhigh) may be better for specific use cases like budget-conscious projects or speed-critical applications.