Open LLM Leaderboard v2
Indexed results from Open LLM Leaderboard v2 · Archive. Each ranking retains its own metric and evaluation conditions.
Open LLM v2 MMLU-PRO · Normalized
openLLMLeaderboardV2 · Open LLM Leaderboard v2 · Normalized score · percent
- MaziyarPanahi/calme-3.2-instruct-78bmaziyarpanahi70.034%
- MaziyarPanahi/calme-3.1-instruct-78bmaziyarpanahi68.722%
- dfurman/CalmeRys-78B-Orpo-v0.1dfurman66.801%
- MaziyarPanahi/calme-2.4-rys-78bmaziyarpanahi66.691%
- glaiveai/Reflection-Llama-3.1-70Bglaiveai59.349%
- arcee-ai/Arcee-Blitzarcee-ai57.262%
- newsbang/Homer-v1.0-Qwen2.5-72Bnewsbang57.17%
- Sakalti/ultiima-72B-v1.5sakalti56.154%
- Qwen/Qwen2.5-72Bqwen55.203%
- Aryanne/QwentileSwaparyanne54.953%
- rombodawg/Rombos-LLM-V2.5-Qwen-72brombodawg54.833%
- rombodawg/Rombos-LLM-V2.5-Qwen-32brombodawg54.621%
Models
LoadingBenchmarks
Explore every benchmark
LoadingThese charts contain model-linked results. The charts above also include unlinked source profiles. Compare scores within each evaluation. Model quality, usage, transparency and system performance measure different things.