Quality scores
How this model compares to the rest of the field on each prompt category, from a pairwise fit over the same prompts. The raw judge score sits underneath each number.
Win rate per category: how often this model beats a field-average model on a prompt from that category. 50% is average, not a failing grade. It is not a percentage of correct answers.
Pricing
What you pay per million tokens when you use this model on Tokonomix, plus an estimate for a typical conversation.
Availability
Availability
How often this model answers when we call it — measured across real API requests and live tests over the last 30 days. This is separate from quality: these numbers only tell you whether the model responds, not how good the answer is.
Last 7 days
100.0%
n=5
Last 30 days
100.0%
n=5
Median response time
625ms
n=5
Based on 8 measurements over the last 30 days.
Technical details
Only live API calls and live-test requests count — internal probes and benchmark runs are excluded.
Calls with a custom API key (BYOK) are excluded: those failures are key-specific, not a sign of model downtime.
Failed calls are NOT included in quality scores — quality is measured on successful responses only. Availability and quality are independent signals.
Median response time (p50) across successful calls with a recorded duration. Outliers (very slow or very fast calls) pull the median less than the average.
Total calls (30d)
5
OK responses (30d)
5
Total calls (7d)
5
OK responses (7d)
5
Tokonomix benchmark verdicts
GPT-5.6-Terra Maintains Stable Performance in Second Benchmark Window
GPT-5.6-Terra shows consistent performance across the second benchmark window with no significant changes detected in its core capabilities. The model continues to deliver the same level of quality that characterized its debut, maintaining stable scores across all evaluation metrics. Since this is only the second benchmark window and the primary detected change involves pricing rather than performance characteristics, users can expect the same capabilities they observed in the initial release. The model's performance profile remains unchanged across reasoning, coding, mathematical, and language understanding tasks. For organizations already using GPT-5.6-Terra, this stability means existing workflows and integrations should continue to perform as expected without degradation or unexpected improvements that might require adjustment. The lack of performance variation between windows suggests reliable and predictable behavior, which can be valuable for production deployments. Users evaluating this model should refer to the initial benchmark results as they remain representative of current capabilities. The Azure OpenAI deployment in the EU Sweden region continues to provide the same service level established in the previous window.
Quality
—
Latency p50
—
Test runs
0