The most notable movement in today's data is a total systemic failure of availability across multiple model families. Every measured model in the speed and uptime category reported a 0% success rate over the last 24 hours.
Speed
Median response times could not be calculated because no requests succeeded. The following models showed 0% uptime across all hourly checks:
| Model | 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
Google Gemini models did not respond during the latest pricing check. The last observed billed costs per million tokens are as follows:
- Gemini 3.5 Flash: $1.50/M in, $9.00/M out
- Gemini 3.1 Flash Lite: $0.250/M in, $1.50/M out
- Gemini 3.1 Pro (preview): $2.00/M in, $12.00/M out
- Gemini 2.5 Flash: $0.300/M in, $2.50/M out
- Gemini 2.5 Flash Lite: $0.100/M in, $0.400/M out
- Gemini 2.5 Pro: $1.25/M in, $10.00/M out
Benchmark
No standing benchmark data was recorded for the last 7 days. Total spend to run benchmarks is not available.
These numbers prove that the tested endpoints were unreachable during the measurement windows. They do not prove the cause of the downtime or whether the models are functional via alternative access points.
