LA Leaderboard - Parafraseja: leaderboard
Paraphrase identification on Catalan sentence pairs (Parafraseja). From the LA Leaderboard, which evaluates models on Spanish, Catalan, Basque and Galician tasks with the lm-evaluation-harness; accuracy (%), higher is better. Each model result file is a separate run.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 69 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen3-8B-Base | 68.3 | #626 |
| 2 | Llama 3.1 8B Instruct | 67.55 | #1017 |
| 3 | Mistral 7B Instruct (v0.3) | 67.1 | #1255 |
| 4 | Gemma 2 9B (IT) | 65.45 | #795 |
| 5 | Qwen 2.5 7B | 65.45 | #895 |
| 6 | occiglot-7B-eu5 | 64.88 | |
| 7 | Mistral-7B-v0.3 | 64.48 | #1321 |
| 8 | gemma-3-4B-pt Base | 64.4 | #1110 |
| 9 | Gemma 2 9B | 64.25 | #924 |
| 10 | Yi-1.5-9B | 64.2 | #1215 |
| 11 | Llama 3.1 8B | 64.12 | #1130 |
| 12 | occiglot-7B-eu5-instruct | 64.05 | |
| 13 | salamandra-7B | 63.65 | #1446 |
| 14 | Teuken-7B-Instruct-v0.4 | 63.6 | #1408 |
| 15 | EuroLLM-9B-Instruct | 63.48 | #1047 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=la-leaderboard-parafraseja · How It Works · Data refreshed daily, snapshot 2026-11-05.