Natural Language Understanding on italic fast
86.1Exact Match (EM)GPT-5 nano
Evaluation results
| Method | Links | |
|---|---|---|
| GPT-5 nanointegration=lm_evaluation_harness2026.05 | 86.1 | |
| Ministral-3-8B-Instruct-2512-BF16integration=lm_evaluation_harness2026.05 | 80.2 | |
| gemma-3-12b-itintegration=lm_evaluation_harness2026.05 | 76.5 | |
| Qwen3-8Bintegration=lm_evaluation_harness2026.05 | 73 | |
| LLaMAntino-3-ANITA-8B-Inst-DPO-ITAintegration=lm_evaluation_harness2026.05 | 69.9 | |
| Llama-3.1-8B-Instructintegration=lm_evaluation_harness2026.05 | 67.5 | |
| Velvet-14Bintegration=lm_evaluation_harness2026.05 | 67.2 | |
| gpt-oss-20bintegration=lm_evaluation_harness2026.05 | 67.2 | |
| Qwen3-4Bintegration=lm_evaluation_harness2026.05 | 66.7 | |
| gemma-3-4b-itintegration=lm_evaluation_harness2026.05 | 62.8 | |
| FastwebMIIA-7Bintegration=lm_evaluation_harness2026.05 | 61.7 | |
| EngGPT2-16B-A3Bintegration=lm_evaluation_harness2026.05 | 60.2 | |
| Llama-3.2-3B-Instructintegration=lm_evaluation_harness2026.05 | 57.4 | |
| Moonlight-16B-A3B-Instructintegration=lm_evaluation_harness2026.05 | 57 | |
| deepseek-moe-16b-chatintegration=lm_evaluation_harness2026.05 | 48.5 | |
| Minerva-7B-instruct-v1.0integration=lm_evaluation_harness2026.05 | 34.7 |