Named Entity Recognition on MultiCoNER v1
42.8Macro F1Gemma3-27B
Evaluation results
| Method | Links | |
|---|---|---|
| Gemma3-27B2026.01 | 42.8 | |
| Qwen3-32B2026.01 | 41.9 | |
| OTTERArchitecture=Cross-Encoder, Backbone=mmBERT, Training Steps=100k, Thresholding Strategy=Best per language2026.01 | 36.9 | |
| GLiNER-multi-v2.12026.01 | 36.6 | |
| OTTERArchitecture=Cross-Encoder, Backbone=mmBERT, Training Steps=100k2026.01 | 35.8 | |
| OTTERArchitecture=Bi-Encoder, Backbone=RemBERT2026.01 | 35.1 | |
| OTTERArchitecture=Cross-Encoder, Backbone=RemBERT2026.01 | 34.7 | |
| OTTERArchitecture=Bi-Encoder, Backbone=mmBERT2026.01 | 34.3 | |
| OTTERArchitecture=Cross-Encoder, Backbone=mmBERT2026.01 | 33.8 | |
| OTTERArchitecture=Bi-Encoder, Backbone=mmBERT, Training Steps=100k2026.01 | 33.5 | |
| GLiNER-x-base2026.01 | 32.9 | |
| GPT-52026.01 | 26.7 | |
| WikiNeural2026.01 | 9.7 |