Language Modeling on Portuguese Evaluation Suite (test)
5.92NPMGlorIA-1.3B
Evaluation results
| Method | Links | |
|---|---|---|
| GlorIA-1.3BParams (B)=1.3, PT Tokens (B)=35, Total Tokens (B)=35, PT FLOPs=2.73e+20, Total FLOPs=2.73e+202026.03 | 5.92 | |
| TeenyTinyLlama-160mParams (B)=0.16, PT Tokens (B)=6.2, Total Tokens (B)=6.2, PT FLOPs=5.95e+18, Total FLOPs=5.95e+182026.03 | 7.71 | |
| Tucano-160mParams (B)=0.16, PT Tokens (B)=169, Total Tokens (B)=169, PT FLOPs=1.62e+20, Total FLOPs=1.62e+202026.03 | 8.78 | |
| TeenyTinyLlama-460mParams (B)=0.46, PT Tokens (B)=6.2, Total Tokens (B)=6.2, PT FLOPs=1.71e+19, Total FLOPs=1.71e+192026.03 | 11.18 | |
| Carvalho-pt-gl-1.3BParams (B)=1.3, PT Tokens (B)=26, CPT Tokens (B)=5, Total Tokens (B)=31, PT FLOPs=2.03e+20, CPT FLOPs=3.90e+19, Total FLOPs=2.42e+202026.03 | 12.54 | |
| Tucano-630mParams (B)=0.63, PT Tokens (B)=211, Total Tokens (B)=211, PT FLOPs=7.98e+20, Total FLOPs=7.98e+202026.03 | 14.89 | |
| Tucano-1b1Params (B)=1.1, PT Tokens (B)=250, Total Tokens (B)=250, PT FLOPs=1.65e+21, Total FLOPs=1.65e+212026.03 | 15.44 | |
| Llama-3.2-1BParams (B)=1.0, PT Tokens (B)=9,000, Total Tokens (B)=9,000, PT FLOPs=5.40e+22, Total FLOPs=5.40e+222026.03 | 16.57 | |
| Curio-Edu-1BParams (B)=1.1, PT Tokens (B)=1,000, CPT Tokens (B)=20, Total Tokens (B)=1,020, PT FLOPs=6.60e+21, CPT FLOPs=1.32e+20, Total FLOPs=6.73e+212026.03 | 17.72 | |
| Tucano-2b4Params (B)=2.4, PT Tokens (B)=515, Total Tokens (B)=515, PT FLOPs=7.42e+21, Total FLOPs=7.42e+212026.03 | 17.87 | |
| Curio-1.1BParams (B)=1.1, PT Tokens (B)=1,000, CPT Tokens (B)=150, Total Tokens (B)=1,150, PT FLOPs=6.60e+21, CPT FLOPs=9.90e+20, Total FLOPs=7.59e+212026.03 | 19.23 | |
| Qwen2.5-0.5BParams (B)=0.5, PT Tokens (B)=18,000, Total Tokens (B)=18,000, PT FLOPs=5.40e+22, Total FLOPs=5.40e+222026.03 | 19.89 | |
| Tucano2-0.6B-BaseParams (B)=0.6, PT Tokens (B)=408, Total Tokens (B)=408, PT FLOPs=1.47e+21, Total FLOPs=1.47e+212026.03 | 20.63 | |
| Llama-2-7BParams (B)=7.0, PT Tokens (B)=2,000, Total Tokens (B)=2,000, PT FLOPs=8.40e+22, Total FLOPs=8.40e+222026.03 | 29.36 | |
| Qwen3-0.6B-BaseParams (B)=0.6, PT Tokens (B)=36,000, Total Tokens (B)=36,000, PT FLOPs=1.30e+23, Total FLOPs=1.30e+232026.03 | 29.39 | |
| Tucano2-qwen-0.5B-BaseParams (B)=0.5, PT Tokens (B)=36,000, CPT Tokens (B)=50, Total Tokens (B)=36,050, PT FLOPs=1.30e+23, CPT FLOPs=1.50e+20, Total FLOPs=1.30e+232026.03 | 35.35 | |
| Granite-3.3-2BParams (B)=2.0, PT Tokens (B)=12,000, Total Tokens (B)=12,000, PT FLOPs=1.44e+23, Total FLOPs=1.44e+232026.03 | 39.96 | |
| Llama-3.2-3BParams (B)=3.0, PT Tokens (B)=9,000, Total Tokens (B)=9,000, PT FLOPs=1.62e+23, Total FLOPs=1.62e+232026.03 | 40.5 | |
| Curio-7BParams (B)=7.0, PT Tokens (B)=2,000, CPT Tokens (B)=150, Total Tokens (B)=2,150, PT FLOPs=8.40e+22, CPT FLOPs=6.30e+21, Total FLOPs=9.03e+222026.03 | 42.78 | |
| Qwen3-1.7B-BaseParams (B)=1.7, PT Tokens (B)=36,000, Total Tokens (B)=36,000, PT FLOPs=3.67e+23, Total FLOPs=3.67e+232026.03 | 44.48 | |
| Curio-Edu-7BParams (B)=7.0, PT Tokens (B)=2,000, CPT Tokens (B)=20, Total Tokens (B)=2,020, PT FLOPs=8.40e+22, CPT FLOPs=8.40e+20, Total FLOPs=8.48e+222026.03 | 45.66 | |
| Tucano2-qwen-1.5B-BaseParams (B)=1.5, PT Tokens (B)=36,000, CPT Tokens (B)=100, Total Tokens (B)=36,100, PT FLOPs=3.67e+23, CPT FLOPs=9.0e+20, Total FLOPs=3.68e+232026.03 | 47.89 | |
| Qwen2.5-3BParams (B)=3.0, PT Tokens (B)=18,000, Total Tokens (B)=18,000, PT FLOPs=3.24e+23, Total FLOPs=3.24e+232026.03 | 50.15 | |
| SmolLM3-3B-BaseParams (B)=3.0, PT Tokens (B)=11,200, Total Tokens (B)=11,200, PT FLOPs=2.02e+23, Total FLOPs=2.02e+232026.03 | 50.25 | |
| Qwen3-4B-BaseParams (B)=4.0, PT Tokens (B)=36,000, Total Tokens (B)=36,000, PT FLOPs=8.64e+23, Total FLOPs=8.64e+232026.03 | 57.86 | |
| Qwen2.5-7BParams (B)=7.0, PT Tokens (B)=18,000, Total Tokens (B)=18,000, PT FLOPs=7.56e+23, Total FLOPs=7.56e+232026.03 | 57.97 | |
| Tucano2-qwen-3.7B-BaseParams (B)=3.7, PT Tokens (B)=36,000, CPT Tokens (B)=50, Total Tokens (B)=36,050, PT FLOPs=8.64e+23, CPT FLOPs=1.11e+21, Total FLOPs=8.65e+232026.03 | 59.21 |