Models
| Model | Params | Context |
|---|---|---|
| Llama 3.1 8B Instruct | 8B | 128K |
| Llama 3.3 70B Instruct | 70.6B | 128K |
| Llama 3.2 3B Instruct | 3.2B | 128K |
| Llama 4 Scout (17B active, 16 experts) | 109B (17B active) | 9,765.625K |
| Qwen3 8B | 8.2B | 32K |
| Qwen3 14B | 14.8B | 32K |
| Qwen3 32B | 32.8B | 32K |
| Qwen3 30B-A3B (MoE) | 30.5B (3.3B active) | 32K |
| Qwen3 235B-A22B (MoE) | 235B (22B active) | 32K |
| Qwen2.5 Coder 32B Instruct | 32.8B | 128K |
| Gemma 3 4B | 4.3B | 128K |
| Gemma 3 12B | 12.2B | 128K |
| Gemma 3 27B | 27.4B | 128K |
| Mistral Small 3.1 24B | 24B | 128K |
| Mistral Nemo 12B | 12.2B | 128K |
| Phi-4 14B | 14.7B | 16K |
| DeepSeek R1 Distill Qwen 32B | 32.8B | 128K |
| DeepSeek R1 Distill Llama 70B | 70.6B | 128K |
| gpt-oss-20b (MoE) | 21B (3.6B active) | 128K |
| gpt-oss-120b (MoE) | 117B (5.1B active) | 128K |
| DeepSeek V3 / R1 671B (MoE) | 671B (37B active) | 128K |
No suitable GPU? Rent one by the hour on Vast.ai or RunPod (affiliate links).