This week's launches: o3, o4-mini, GPT-4.1, and Codex CLI

Announcements
3

Variance in 1st token time and total 256 token response time across 100 trials.

Note that each graph is independently scaled, and scaling to 12 bins with non-zero values.

gpt-4.1-mini time to first token

A histogram showing the distribution of default model latency in seconds, with most values concentrated below 1 second. (Captioned by AI)
A histogram showing the distribution of default model latency in seconds, with most values concentrated below 1 second. (Captioned by AI)1389×989 26.3 KB

gpt-4.1-nano time to first token

The histogram illustrates the latency distribution of GPT-4-Nano, with most occurrences clustered between 0.2 and 0.5 seconds. (Captioned by AI)
The histogram illustrates the latency distribution of GPT-4-Nano, with most occurrences clustered between 0.2 and 0.5 seconds. (Captioned by AI)1389×989 34.7 KB

gpt-4.1-mini total time to 256 tokens

This histogram displays the distribution of total time taken by a default model, with most instances clustered between 1 and 3 seconds. (Captioned by AI)
This histogram displays the distribution of total time taken by a default model, with most instances clustered between 1 and 3 seconds. (Captioned by AI)1389×989 32.8 KB

gpt-4.1-nano total time to 256 tokens

This histogram shows the distribution of "GPT-4-Nano Total Time" with the majority of counts concentrated between 0 and 2 seconds. (Captioned by AI)
This histogram shows the distribution of "GPT-4-Nano Total Time" with the majority of counts concentrated between 0 and 2 seconds. (Captioned by AI)1389×989 24.9 KB

No retries tolerated, all success. Cache-breaking patterns with 1800 tokens in. Launch rate: 600 RPM

Metric Average Minimum Maximum
mini total TPS 94.1 26.8 128.6
nano total TPS 153.1 26.1 242.7

nano: 256 tokens in under two seconds in over 60%.

1 Like