AI · Technology · 2026-09-09 09:30:00
This tracker detects statistically significant degradations in Codex with gpt-5.6-sol performance on SWE tasks. Status is nominal, with no time period showing a statistically significant drop (p < 0.05). The baseline pass rate is 83 percent. Today's pass rate is 86 percent across 50 eval test cases; the past 7-day rate is 85 percent across 350 cases; the past 30-day rate is 85 percent across 1,500 cases. It also tracks daily input tokens, output tokens, average runtime, and tool calls.
Auto-generated summary for reference only. RSS