Tier list
Local AI model tier list
Every model we track, ranked by quality alone. Tiers don’t depend on your hardware: they tell you how capable a model is. To see which ones your machine can run, use the Model Finder.
Step 1
Tell us your computer first
Then you will see how many of these 63 decision models run on it, and can filter the list by how well each one runs.
S
Excellent · 85+ of 100
A
Great · 70+ of 100
Show or hide the list
- Granite Guardian 4.1 8B
- Jina Reranker v3
- Llama 3.1 Nemotron Safety Guard 8B v3
- Qwen3Guard Gen 4B
- Skywork Reward V2 Qwen3 8B
- mxbai-rerank-large v2
- GLiNER2.5 Decide
- gpt-oss-safeguard 120B
- Granite Guardian 3.3 8B
- Qwen3 Reranker 0.6B
- Selene 1 Mini 8B
- WildGuard 7B
- zerank-1 small
- DynaGuard 8B
- BGE Reranker v2 Gemma
- CompassJudger-2 7B
- Nemotron Content Safety Reasoning 4B
- OpenAI Privacy Filter
- gpt-oss-safeguard 20B
- Llama Prompt Guard 2 86M
- mxbai-rerank-base v2
- NuExtract 2.0 4B
- GLiClass ModernBERT Large v3
- Prometheus 2 8x7B
- Qwen3Guard Gen 0.6B
- BGE Reranker v2 M3
- DeBERTa v3 Large Zero-shot v2.0
- DuoGuard 0.5B
- Skywork Reward V2 Qwen3 0.6B
- DynaGuard 1.7B
- GTE Reranker ModernBERT Base
- Jina Reranker v2 Base Multilingual
- ModernBERT Large Zero-shot v2.0
- NuExtract 2.0 2B
B
Good · 55+ of 100
Show the list
- GLiGuard 300M
- GLiNER2 Base v1
- Granite Embedding Reranker English R2
- Llama Guard 3 8B
- Llama Prompt Guard 2 22M
- Prometheus 2 7B
- GLiClass Base v3
- GLIDER
- LettuceDetect Large
- Llama Guard 4 12B
- ProtectAI DeBERTa v3 Prompt Injection v2
- ShieldGemma 27B
- DeBERTa v3 Base Zero-shot v2.0
- Flow Judge v0.1
- ShieldGemma 9B
- GLiNER Multi v2.1
- HHEM 2.1 Open
- ShieldGemma 2B
- BART Large MNLI
- Llama Guard 3 1B
C
Fair · 40+ of 100
Show the list
Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.