Meerada Grade — Live Board
Continuously measured on verifiable tasks · quality & cost per accepted task · every grade shows its confidence · 2026-10-26 17:48Z
| # | Model | Grade | Quality (Wilson 95%) | n | CPAT | TTAT | Confidence |
|---|
| 🥇 |
groq/compound |
97 |
96.7% [83-99] |
30 |
$0.00100 ~ |
5.3s |
provisional |
| 🥈 |
groq/compound-mini |
88 |
88.2% [73-95] |
34 |
$0.00061 ~ |
2.4s |
provisional |
| 🥉 |
openai/gpt-oss-120b |
71 |
71.4% [61-80] |
77 |
$0.00028 |
1.1s |
provisional |
| 4 |
qwen/qwen3.8-27b |
71 |
70.6% [54-83] |
34 |
$0.00052 |
0.6s |
provisional |
| 5 |
openai/gpt-oss-20b |
58 |
57.7% [46-69] |
71 |
$0.00006 |
1.2s |
provisional |
| 6 |
allam-2-7b |
39 |
39.2% [29-51] |
74 |
$0.00009 ~ |
0.6s |
provisional |
| 7 |
qwen/qwen3.6-27b |
31 |
30.6% [18-47] |
36 |
$0.00202 |
45.9s |
provisional |