​
Supported Models

The following open source models are supported. View the public model list here

You can use the following model name as model_id in Heurist API/SDK.

​
Large Language Models (LLMs)

  • deepseek/deepseek-r1: DeepSeek R1 is a groundbreaking open-source AI model that achieves performance comparable to OpenAI’s o1 model across math, code, and reasoning tasks, supporting self-verification and reflection, while being more cost-efficient than its competitors.
  • deepseek/deepseek-v3: DeepSeek V3 0324 version is a powerful Mixture-of-Experts (MoE) language model with 685B total parameters, activating 37B parameters for each token. It demonstrates notable improvements over its predecessor, DeepSeek-V3, and achieves results comparable to Claude Sonnet 3.7.
  • deepseek/deepseek-r1-distill-llama-70b: A distilled version of DeepSeek R1, using Llama 3 70B as the base model. It achieves results comparable to DeepSeek R1 but much more cost-efficient and faster.
  • NaniDAO/deepseek-r1-qwen-2.5-32B-ablated: A distilled version of DeepSeek R1, with using Qwen 2.5 32B as the base model. It uses ablation technique to remove refusal mechanism to make the model always generate responses. Uncensored.
  • mistralai/mixtral-8x22b-instruct: A Sparse Mixture-of-Experts model featuring 8 experts and 22 billion parameters, optimized for instruction-following with strong performance in reasoning and multilingual capabilities.
  • mistralai/mixtral-8x7b-instruct: A compact version of the Mixtral family, designed for efficient instruction-based interactions while maintaining a balance between performance and cost.
  • mistralai/mistral-7b-instruct: Excel in generating fast responses and highly cost-efficient.
  • meta-llama/llama-3.3-70b-instruct: The latest Llama 3 model, outperforming many of the available open source and closed chat models on common industry benchmarks.
  • nvidia/llama-3.1-nemotron-70b-instruct: specialized version of the Llama model tailored by NVIDIA for complex instruction-following tasks, delivering high-quality, human-like responses across a variety of applications while leveraging advanced NVIDIA AI technologies for optimal performance and scalability.
  • qwen/qwen-2.5-coder-32b-instruct: State-of-the-art open-source code model, matching the coding capabilities of GPT-4o.
  • hermes-3-llama3.1-8b: Flagship Hermes LLM trained by Nous Research, with advanced agentic capabilities, enhanced roleplaying, reasoning, multi-turn conversation, long context coherence and agentic abilities. Uncensored. Supports tool calling
  • qwen/qwen-2.5-coder-32b-instruct: Qwen2.5-Coder is the latest series of Code-Specific Qwen large language models (formerly known as CodeQwen)
  • asi1-mini: ASI1-mini is the first Web3-native LLM, specifically built and optimized for supporting complex agentic workflows. Developed by Fetch.ai, it features adaptive reasoning and context-aware decision-making.

​
Note on uncensored models:

The models marked as Uncensored are finetuned with specific datasets to eliminate censorship. For other models, it’s possible to avoid censorship by using jailbreaking prompts, which works in most cases.

​
Image Generation Models

  • HeuristLogo: Flux LoRA that can generate the logo of Heurist. Trigger word: Heuristai logo or hexagonal logo.

  • FLUX.1-dev: State-of-the-art open source image generation model that excels at a variety of image styles.

  • PepeXL: SDXL LoRA that can generate the meme frog, Pepe. Trigger word: Pepe. Recommended words: pepe_frog, meme, web comic and add cartoon or 3d render

  • Zeek: SDXL LoRA for ZKsync’s mascot cat, Zeek. Trigger word: zeek. We recommend using zeek wearing ... or zeek disguised as ...

  • Aurora: SD1.5 checkpoint for anime girls.

  • AnimagineXL: SDXL checkpoint for anime images. It can generate characters from well-known anime series.

  • CyberRealisticXL: SDXL checkpoint for realistic portraits.

  • BrainDance: SD1.5 checkpoint for cartoon, anime and watercolor styles.

  • BlazingDrive: SD1.5 checkpoint for anime characters and scenes.

  • YamersCartoonArcadia: SD1.5 checkpoint for stylized 2D cartoon.

  • ArchitectureRealMix: SD1.5 checkpoint for architectural design, landscape design, urban design, and interior design.

  • ArthemyComics: SD1.5 checkpoint for fantasy cartoon images.

  • ArthemyReal: SD1.5 checkpoint for fantasy-style and realistic-looking images.

  • DreamShaperXL: SDXL checkpoint that aims at doing everything well, including photos, art, anime, and manga styles.

  • AAMXLAnimeMix: SDXL checkpoint for anime art and hentai.

  • SDXLUnstableDiffusersV11: SDXL checkpoint that enhances SDXL capabilities in creating vibrant arts, designs and photo-realistic images.

  • SDXL: General-purpose image generation model developed by Stability AI.

​
Embedding Models

  • BAAI/bge-large-en-v1.5: A large-scale multilingual embedding model. It converts text into a 1024 dimensional vector representation. The max input length is 512 tokens.

​
Need More Models?

Any image generation models from Civitai and LLMs from HuggingFace can be supported upon request. If you’re interested in hosting your models on Heurist, or if you want customized models that adapts to your specific use cases, please contact us at team@heurist.xyz