selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

Alternatives to DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic

120 open-weight models that do the same job as DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic (9.4B chat), ranked by how close they are in task and size, then by popularity and how recent they are. Licence and VRAM are shown for each.

the original

A compact 9B reasoning model distilled from DeepSeek-V4-Pro in Max Effect mode

9.4B256K ctx Apache 2.0
27KOct 2026gguf
--vram
any
--gpus
--quant
  1. tokyotech-llm logo≈6.3 GB

    Llama 3.1 Swallow is a series of large language models (8B, 70B) that were built by continual pre-training on the Meta Llama 3.1 models. Llama 3.1 Swallow enhanced the Japanese language capabilities of the original Llama 3.1 while retaining the English language capabilities.

    8B128K ctx Llama 3.3 Community License
    6.9KJun 2025gguf
    • Same category
    • Similar size (8B vs 9.4B)
    vs
  2. microsoft logo

    UserLM-8b

    microsoft·Chat

    ≈6.3 GB

    Unlike typical LLMs that are trained to play the role of the "assistant" in conversation, we trained UserLM-8b to simulate the “user” role in conversation (by training it to predict user turns in a large corpus of conversations called WildChat).

    8B8K ctx MIT
    2.8KSept 2025gguf
    • Same category
    • Similar size (8B vs 9.4B)
    • Commercial use OK
    vs
  3. allenai logo≈8.8 GB

    We introduce Olmo 3, a new family of 7B and 32B models both Instruct and Think variants. Long chain-of-thought thinking improves reasoning tasks like math and coding.

    7.3B64K ctx Apache 2.0
    4KNov 2025gguf
    • Same category
    • Similar size (7.3B vs 9.4B)
    • Commercial use OK
    vs
  4. inclusionAI logo

    LLaDA2.1-mini

    inclusionAI·Chat

    ≈10 GB

    🚀 LLaDA2.1-flash is now live on ZenmuxAI! Try it via API 🛠️ or Chat 💬: https://zenmux.ai/inclusionai/llada2.1-flash

    16.3BA1.4B32K ctx Apache 2.0
    28.2KFeb 2026
    • Same category
    • Similar size (16.3B vs 9.4B)
    • Commercial use OK
    vs
  5. nvidia logo≈6.5 GB

    We're excited to introduce Nemotron-Cascade-8B, a powerful general-purpose model trained through sequential and domain-wise reinforcement learning. Nemotron-Cascade-8B is post-trained from the Qwen3-8B-Base model.

    8.2B32K ctx NVIDIA Open Model License
    1.1KDec 2025gguf
    • Same category
    • Similar size (8.2B vs 9.4B)
    • Commercial use OK
    vs
  6. Qwen logo≈11 GB

    Qwen2.5 is the latest series of Qwen large language models. For Qwen2.5, we release a number of base language models and instruction-tuned language models ranging from 0.5 to 72 billion parameters. Qwen2.5 brings the following improvements upon Qwen2:

    14.8B32K ctx Apache 2.0
    603.8KSept 2024ollama
    • Same category
    • Similar size (14.8B vs 9.4B)
    • Commercial use OK
    vs
  7. tiiuae logo

    Falcon-H1R-7B

    tiiuae·Chat

    ≈5.3 GB

    This repository presents Falcon-H1R-7B, a reasoning-specialized model introduced in the paper Falcon-H1R: Pushing the Reasoning Frontiers with a Hybrid Model for Efficient Test-Time Scaling.

    7.6B256K ctx Falcon LLM License
    3.2KOct 2025gguf
    • Same category
    • Similar size (7.6B vs 9.4B)
    vs
  8. zai-org logo

    glm-4-9b

    zai-org·Chat

    ≈6.4 GB

    glm-4-9b is an open-weight chat model with 9.4B parameters from zai-org, published on Hugging Face.

    9.4B8K ctx GLM-4 License
    22.9KJun 2024gguf
    • Same category
    • Similar size (9.4B vs 9.4B)
    vs
  9. zai-org logo

    GLM-Z1-9B-0414

    zai-org·Chat

    ≈6.4 GB

    The GLM family welcomes a new generation of open-source models, the GLM-4-32B-0414 series, featuring 32 billion parameters. Its performance is comparable to OpenAI's GPT series and DeepSeek's V3/R1 series, and it supports very user-friendly local deployment features.

    9.4B32K ctx MIT
    2.2KApr 2025gguf
    • Same category
    • Similar size (9.4B vs 9.4B)
    • Commercial use OK
    vs
  10. allenai logo

    wildguard

    allenai·Chat

    ≈5.8 GB

    WildGuard is an open one-stop moderation model that achieves three goals: 1) Detection of harm in users prompts. 2) Detection of harm in LLMs responses. 3) Refusal evaluation of LLMs responses.

    7.2BApache 2.0
    190.7KJun 2024
    • Same category
    • Similar size (7.2B vs 9.4B)
    • Commercial use OK
    vs
  11. Nanbeige logo

    Nanbeige4.2-3B

    Nanbeige·Chat

    ≈3.7 GB

    Nanbeige4.2-3B is a compact agentic model built on Nanbeige4.2-3B-Base, designed to combine strong agentic behavior with broad reasoning and alignment capabilities. Its Looped Transformer architecture reuses the transformer layers to increase model capacity without adding parameters.

    4.2B256K ctx Apache 2.0
    89.1KJul 2026gguf
    • Same category
    • Smaller (4.2B vs 9.4B)
    • Commercial use OK
    vs
  12. nvidia logo≈6.3 GB

    Llama-3.1-Nemotron-Nano-8B-v1 is a large language model (LLM) which is a derivative of Meta Llama-3.1-8B-Instruct (AKA the reference model). It is a reasoning model that is post trained for reasoning, human chat preferences, and tasks, such as RAG and tool calling.

    8B128K ctx NVIDIA Open Model License
    9.8KMar 2025gguf
    • Same category
    • Similar size (8B vs 9.4B)
    • Commercial use OK
    vs
  13. tencent logo

    ContextPilot-14B

    tencent·Chat

    ≈11 GB

    ContextPilot-14B is the Qwen3-14B checkpoint of ContextPilot, a proactive context-management framework for long-horizon language-model agents. It teaches agents to plan, maintain long-term memory, and offload less useful context while they continue reasoning and using tools.

    14.8B40K ctx Custom licence
    2.4KAug 2026gguf
    • Same category
    • Similar size (14.8B vs 9.4B)
    vs
  14. zai-org logo

    glm-4-9b-chat-hf

    zai-org·Chat

    ≈6.4 GB

    If you are using the weights from this repository, please update to

    9.4B128K ctx GLM-4 License
    7KOct 2024gguf
    • Same category
    • Similar size (9.4B vs 9.4B)
    vs
  15. allenai logo≈6.8 GB

    Upon the initial release of OLMo-2 models, we realized the post-trained models did not share the pre-tokenization logic that the base models use. As a result, we have trained new post-trained models.

    7.3B4K ctx Apache 2.0
    42.2KDec 2024ollama
    • Same category
    • Similar size (7.3B vs 9.4B)
    • Commercial use OK
    vs
  16. google logo

    gemma-2-9b

    google·Chat

    ≈8.6 GB

    Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models.

    9.2B8K ctx Gemma Terms of Use
    18.9KJun 2024gguf
    • Same category
    • Similar size (9.2B vs 9.4B)
    vs
  17. ibm-granite logo

    granite-guardian-3.3-8b

    ibm-granite·Chat

    ≈6.6 GB

    Model Summary: Granite Guardian 3.3 8b is a specialized Granite 3.3 8B model designed to judge if the input prompts and the output responses of an LLM based system meet specified criteria.

    8.2B128K ctx Apache 2.0
    4.1KJun 2025gguf
    • Same category
    • Similar size (8.2B vs 9.4B)
    • Commercial use OK
    vs
  18. CohereLabs logo

    aya-expanse-8b

    CohereLabs·Chat

    ≈6.3 GB

    Aya Expanse 8B is an open-weight research release of a model with highly advanced multilingual capabilities. It focuses on pairing a highly performant pre-trained Command family of models with the result of a year’s dedicated research from Cohere Labs, including data arbitrage, multilingual…

    8BCC BY-NC 4.0
    25.4KOct 2024ollama
    • Same category
    • Similar size (8B vs 9.4B)
    vs
  19. meta-llama logo

    Meta-Llama-3-8B

    meta-llama·Chat

    ≈6.3 GB

    Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes.

    8B8K ctx Llama 3 Community License
    101.3KApr 2024gguf
    • Same category
    • Similar size (8B vs 9.4B)
    vs
  20. microsoft logo

    phi-4

    microsoft·Chat

    ≈11 GB

    Our training data is an extension of the data used for Phi-3 and includes a wide variety of sources from:

    14.7B16K ctx MIT
    218.6KDec 2024ollama
    • Same category
    • Similar size (14.7B vs 9.4B)
    • Commercial use OK
    vs
  21. Qwen logo

    Qwen3-4B

    Qwen·Chat

    ≈4.0 GB

    Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models.

    4B40K ctx Apache 2.0
    3.1MApr 2025ollama
    • Same category
    • Smaller (4B vs 9.4B)
    • Commercial use OK
    • Widely used
    vs
  22. tencent logo

    Hunyuan-7B-Instruct

    tencent·Chat

    ≈6.0 GB

    Hunyuan is Tencent's open-source efficient large language model series, designed for versatile deployment across diverse computational environments.

    7.5B32K ctx Unknown licence
    5.4KJul 2025gguf
    • Same category
    • Similar size (7.5B vs 9.4B)
    vs
  23. allenai logo≈8.8 GB

    We introduce Olmo 3, a new family of 7B and 32B models both Instruct and Think variants. Long chain-of-thought thinking improves reasoning tasks like math and coding.

    7.3B64K ctx Apache 2.0
    2.4KDec 2025gguf
    • Same category
    • Similar size (7.3B vs 9.4B)
    • Commercial use OK
    vs
  24. EssentialAI logo

    rnj-1

    EssentialAI·Chat

    ≈6.4 GB

    Rnj-1 is a family of 8B parameter open-weight, dense models trained from scratch by Essential AI, optimized for code and STEM with capabilities on par with SOTA open-weight models.

    8.3B32K ctx Apache 2.0
    884Nov 2025gguf
    • Same category
    • Similar size (8.3B vs 9.4B)
    • Commercial use OK
    vs