Live Benchmark Arena

Qorinix Arena

Race Qorinix 3.1 and Qorinix 3.2 against ChatGPT, DeepSeek, Gemini, and Claude. The live dashboard tracks TTFT, total time, throughput, and output size while exact route internals, credentials, and resilience logic stay server-side.

Public Arena mode: named comparison models are visible. Qorinix routing controls, credentials, API routes, and resilience map stay server-side.

Preset example The Prompt and System Prompt below are prefilled examples for a quick benchmark. You can run them as-is, edit either field, or replace both with your own test case.

Prompt

System Prompt

Profile

Max Output Tokens

Enabled lanes

Free preview: 5 runs remaining.

Ready.

Winner: -
Best TTFT: -
Total: -
Progress: 0/6

Rank	Lane	Status	TTFT	Total	Tok/s

Live Benchmark Arena

Comparison Dashboard

Watch all six lanes race the same prompt in real time. Cards highlight the current leader (gold flag) and pin Qorinix lanes (orange tag) so the speed and cost gap against comparison models stays obvious.

Ready.

TTFT first token delay, lower is better

Total Latency end-to-end time, lower is better

Throughput generation speed, higher is better

Output Size words vs estimated tokens

Benchmark Signals

P1 Commercial Benchmark Layer

Single-run metrics update live. Percentiles, value score, and reliability signals are shown as current-run proxies until historical Arena storage is activated.