The primary finding for October 24 is a total loss of connectivity across all measured high-end and lite models from Google and OpenAI. No median response times were recorded as 0% of hourly checks succeeded for the sampled models.
Speed
Availability collapsed for both the GPT-5 and Gemini series. The following table details the failure rate across the last 24 hours:
| Model | Median Response | Uptime |
|---|---|---|
| GPT-5.4 Mini | — | 0% |
| GPT-5.6 Sol | — | 0% |
| GPT-5.6 Terra | — | 0% |
| GPT-5.6 Luna | — | 0% |
| Gemini 3.1 Pro (preview) | — | 0% |
| Gemini 3.1 Flash Lite | — | 0% |
| Gemini 3.8 Flash | — | 0% |
Prices
Despite the lack of responsiveness during checks, the observed billed costs per million tokens for the Gemini family are as follows:
- Gemini 3.5 Flash: $1.5/M in, $9.0/M out
- Gemini 3.1 Flash Lite: $0.250/M in, $1.5/M out
- Gemini 3.1 Pro (preview): $2.0/M in, $12.0/M out
- Gemini 2.5 Flash: $0.300/M in, $2.5/M out
- Gemini 2.5 Flash Lite: $0.100/M in, $0.400/M out
- Gemini 2.5 Pro: $1.25/M in, $10.0/M out
Benchmark
No benchmark data was recorded for the last 7 days. No available points were captured and no spend was logged for the period.
These measurements prove a total absence of successful responses across the tested endpoints during the 24-hour window. They do not prove the cause of the outages or the current status of the models outside of the specific hourly checks performed.
