| description |
|---|
At Shapes.inc, we offer a variety of models for running your shapes. Each model has its own strengths and characteristics. Let's break down the options: |
Models available:
| HUMAN-LIKE | |
|---|---|
| Gemini 2.0 Flash | Good word choice, somewhat resembles Sonnet 3.5 (older version), good memory and knowledge retrieval, commands execution, excellent, creative, intelligent in chit-chatting environment |
| Gemma 2 27b | Very fast, follows presets well, handles complex tasks with concise prompts, somewhat maintains character, displays quirky behavior and humor |
| Gemma 2 9b (Turbo) | Extremely fast natural chatter |
| Llama 3.1 8b (Turbo) | Excellent at realistic and human-like behavior, fastest Llama engine. |
| Llama 3.3 70b (Turbo) | Good for realistic and human-like behavior, with faster performance. |
| QwQ 32b | Smart engine, Great at reasoning! good for versatile tasks. |
| Qwen 2.5 Coder 32b | Another smart engine, Great at reasoning! good overall, especially for coding and maths. |
| Gemini 1.5 Flash | Average overall, good at human-like behavior. |
| Llama 3 70b instruct | Formerly known as llama 3 instruct model. Good at retrieving knowledge, being in character, human-like behavior. |
| INTELLIGENT MODELS | |
| 3.5 Sonnet | Great for roleplay and advanced tasks |
| Gemini 1.5 Pro | Solid all around with roleplay, humanlike speech and tasks |
| Llama 3.1 70b | Good at realistic and human-like behavior and tasks |
| Llama 3.1 70b (Turbo) | Best for realistic and human-like behavior, with faster performance doing tasks |
| DeepSeek V3 | Similar to Sonnet and Gemini it is good for both roleplay and following advanced instructions. |
| DeepSeek V3 R1 | An even more intelligent version of V3 that you can enable to see its reasoning. Great all around model for anything including RP. |
| Deepseek R1 Llama 3.3 70b Distill | |
| ROLEPLAY SPECIFIC MODELS | |
| WizardLM 8x22 | Sends long creative roleplay |
| Mistral Nemo | Short creative role-player and great chatter. Needs temperature between 0.8-0.9 |
| 3.5 Haiku | |
| DEPRECIATED MODELS | |
| WizardLM 2 7b (Turbo) | Fast long role-player but doesn't follow format well. |
| lzlv 70b | (This model is currently being redirected to Llama 3.3 70b Instruct) |
| Lunaris L3 8B | Good for short fast basic roleplay |
Feel free to experiment and find the best engine for your shape!