openrouter.ai · openai ·
openai/gpt-6.1-terra-pro on openrouter.ai
significant issues7 of 7 cases passed
52/100trust score
Observed facts across five dimensions
- Protocol compatibilityAPI format: OpenAI-compatible · tool calls: normal · parameter check: passed (a request with tools came back normally)
- Agent usability7/7 cases completed
- Usage reportingThe relay reported token usage
- Performancetotal duration 72.7 s
- Difference from official baselineduration +92%; input tokens +253%; output tokens +225%
7 cases
| Case | Result | Latency | Baseline | Difference |
|---|---|---|---|---|
| Basic Tool I/O | Passed | 13.7 s | 8.6 s | +60% |
| Multi-File Synthesis | Passed | 22.9 s | 11.9 s | +92% |
| Model Context Synthesis | Passed | 3.7 s | 1.7 s | +119% |
| Single Tool Call | Passed | 6.9 s | 3.8 s | +79% |
| Chained Tool Calls | Passed | 9.6 s | 4.8 s | +100% |
| Error Retry & Recovery | Passed | 10.1 s | 3.9 s | +155% |
| Long-Context Recall | Passed | 2.9 s | 1.5 s | +91% |
Tokens
| Metric | Relay | Baseline | Difference |
|---|---|---|---|
| Input tokens | 189,226 | 53,632 | +253% |
| Output tokens | 2,513 | 774 | +225% |
Previous tests
The 0 tests before the latest.
| Date | Score | Passed |
|---|
How is this relay doing now? Test it with your key
Last tested 2026-10-29 (1 run). Relays change; a test shows how this one does right now (usually 1–3 minutes; the key is never stored).
Test it →The address is prefilled with the usual /v1. Change it if your relay uses another path.