grimjim: AI model benchmarks
AI models from grimjim. 3 models tracked, 3 with a unified ELO rating. Current best: Llama-3.1-8B-Instruct-abliterated_via_adapter (ELO 1472, rank #867 overall). Most recent tracked release: 2024-07-25.
Rated models
| Rank | Model | Provider | Unified ELO | Benchmarks |
|---|---|---|---|---|
| 867 | Llama-3.1-8B-Instruct-abliterated_via_adapter | grimjim | 1472 ± 20 | 21 |
| 1034 | Llama-3-Instruct-8B-SimPO-SPPO-Iter3-merge | grimjim | 1439 ± 17 | 29 |
| 1062 | Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge | grimjim | 1435 ± 17 | 29 |
Interactive version: theaggregate.ai/company?slug=grimjim · How It Works · Data refreshed daily, snapshot 2026-10-07.