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 Arete-27B

120 open-weight models that do the same job as Arete-27B (27.4B 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
NousResearch logo

Arete-27B

NousResearch·Chat

≈19 GB

We post-trained over Alibaba's Qwen3.8-27B to reinforce western values on a Chinese-trained model. We call this new model Arete-27B.

27.4B256K ctx Apache 2.0
7.5KNov 2026gguf
--vram
any
--gpus
--quant
  1. XingChen-AGI logo

    Xing4.0-29B-Intent

    XingChen-AGI·Chat

    ≈27 GB

    For full project documentation, deployment instructions, SDK usage, and data details, visit the GitHub repository.

    31.2BA5.5B256K ctx Apache 2.0
    6.1KOct 2026gguf
    • Same category
    • Similar size (31.2B vs 27.4B)
    • Commercial use OK
    vs
  2. ibm-granite logo

    granite-4.1-30b

    ibm-granite·Chat

    ≈20 GB

    Model Summary: Granite-4.1-30B is a 30B parameter long-context instruct model finetuned from Granite-4.1-30B-Base using a combination of open source instruction datasets with permissive license and internally collected synthetic datasets.

    28.9B128K ctx Apache 2.0
    11.9KApr 2026ollama
    • Same category
    • Similar size (28.9B vs 27.4B)
    • Commercial use OK
    vs
  3. summerAIO logo≈19 GB

    Qwen3.8-27B-Longbook27-v2 is a fine-tuned 27B model for understanding and generating long-form text such as books, papers, and lengthy reports. Version 2 improves on v1 in behavior and output quality.

    27.1B256K ctx Apache 2.0
    1.5KOct 2026
    • Same category
    • Similar size (27.1B vs 27.4B)
    • Commercial use OK
    vs
  4. XingChen-AGI logo

    Xing4.0-29B-A4B

    XingChen-AGI·Chat

    ≈27 GB

    Xing4.0-29B-A4B is a next-generation large language model in the Xing series (formerly TeleChat), developed by China Telecom Artificial Intelligence Technology Co., Ltd. With 29B total parameters and only 4B activated per token, it natively supports a 256K context length, extensible to 512K.

    31.2BA4B256K ctx Apache 2.0
    5.4KSept 2026gguf
    • Same category
    • Similar size (31.2B vs 27.4B)
    • Commercial use OK
    vs
  5. Looking for the post-trained model? See NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 for the full-precision release, or NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4 for optimized inference.

    31.6BA3B256K ctx openmdw-1.1
    5.6KAug 2026gguf
    • Same category
    • Similar size (31.6B vs 27.4B)
    • Commercial use OK
    vs
  6. Qwen logo

    Qwen3-14B

    Qwen·Chat

    ≈11 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.

    14.8B40K ctx Apache 2.0
    13MApr 2025ollama
    • Same category
    • Similar size (14.8B vs 27.4B)
    • Commercial use OK
    • Widely used
    vs
  7. dphn logo≈23 GB

    This is our most spectacular outcome ever. FFT, all parameters, 16bit. 77.4 MMLU on 34b. And it talks like a dream.

    34.4B8K ctx Apache 2.0
    4.8MMay 2024gguf
    • Same category
    • Similar size (34.4B vs 27.4B)
    • Commercial use OK
    • Widely used
    vs
  8. LiquidAI logo

    LFM2.5-24B-A2B-Base

    LiquidAI·Chat

    ≈15 GB

    LFM2.5 is a new family of hybrid models designed for on-device deployment. LFM2.5-24B-A2B is the largest model in the family: a 24B-parameter mixture of experts with only 2B active parameters.

    24BA2B128K ctx lfm1.0
    2.5KOct 2026
    • Same category
    • Similar size (24B vs 27.4B)
    vs
  9. arcee-ai logo

    Trinity-Mini

    arcee-ai·Chat

    ≈16 GB

    Trinity Mini is an Arcee AI 26B MoE model with 3B active parameters. It is the medium-sized model in our new Trinity family, a series of open-weight models for enterprise and tinkerers alike.

    26.1BA3.5B128K ctx openmdw-1.1
    11.8KDec 2025gguf
    • Same category
    • Similar size (26.1B vs 27.4B)
    • Commercial use OK
    vs
  10. IFM logo≈23 GB

    K2-Horizon-32B is the large dense member of the K2-Horizon family: a 32B decoder-only model with a 512K context window.

    34.8B512K ctx Apache 2.0
    8.6KSept 2026gguf
    • Same category
    • Similar size (34.8B vs 27.4B)
    • Commercial use OK
    vs
  11. IFM logo≈24 GB

    K2-Horizon-MoVA-36B-A4B is the sparse member of the K2-Horizon family: a Mixture-of-Experts model with Mixture-of-Values attention (MoVA) that stores 36B parameters and runs 4B per token. We have released the checkpoints, along with the data and the training code.

    37.4BA4B512K ctx Apache 2.0
    14.7KSept 2026gguf
    • Same category
    • Similar size (37.4B vs 27.4B)
    • Commercial use OK
    vs
  12. NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16 is a base large language model (LLM) trained from scratch by NVIDIA, with the next token prediction loss. It provides a good starting point for instruction fine-tuning.

    31.6BA3Bnvidia-nemotron-open-model-license
    18.9KDec 2025gguf
    • Same category
    • Similar size (31.6B vs 27.4B)
    • Commercial use OK
    vs
  13. Qwen logo≈22 GB

    Qwen-AgentWorld is the first language world model to cover seven agent interaction domains within a single model. It simulates agentic environments via long chain-of-thought reasoning, predicting the next environment state given an agent's action and interaction history.

    34.7BA3B256K ctx Apache 2.0
    9.5KJun 2026gguf
    • Same category
    • Similar size (34.7B vs 27.4B)
    • Commercial use OK
    vs
  14. allenai logo≈22 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.

    32.2B64K ctx Apache 2.0
    21.2KDec 2025ollama
    • Same category
    • Similar size (32.2B vs 27.4B)
    • Commercial use OK
    vs
  15. Qwen logo≈22 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:

    32.8B32K ctx Apache 2.0
    718.6KSept 2024ollama
    • Same category
    • Similar size (32.8B vs 27.4B)
    • Commercial use OK
    vs
  16. Nanbeige logo

    Nanbeige4.3-33B-Base

    Nanbeige·Chat

    ≈22 GB

    In Nanbeige4.3, we take a foundational step toward pushing the boundaries of lightweight agents. Our strategy centers on building a high-capacity teacher agent first, which will subsequently anchor the distillation of advanced agentic capabilities into smaller models.

    33.1B128K ctx Apache 2.0
    2.2KOct 2026gguf
    • Same category
    • Similar size (33.1B vs 27.4B)
    • Commercial use OK
    vs
  17. nvidia logo≈20 GB

    We're excited to introduce Nemotron-Cascade-2-30B-A3B, an open 30B MoE model with 3B activated parameters that delivers strong reasoning and agentic capabilities. It is post-trained from the Nemotron-3-Nano-30B-A3B-Base.

    31.6BA3B256K ctx NVIDIA Open Model License
    7KMar 2026ollama
    • Same category
    • Similar size (31.6B vs 27.4B)
    • Commercial use OK
    vs
  18. InternScience logo

    MLExpert

    InternScience·Chat

    ≈22 GB

    MLExpert is a 35B Mixture-of-Experts agentic model for autonomous machine learning engineering. It organizes experiments, writes and revises code, and improves solutions through sustained interaction with execution environments.

    34.7BA3.5B256K ctx Apache 2.0
    3.4KOct 2026gguf
    • Same category
    • Similar size (34.7B vs 27.4B)
    • Commercial use OK
    vs
  19. ibm-granite logo

    granite-4.0-h-small

    ibm-granite·Chat

    ≈21 GB

    📣 Update [10-07-2025]: Added a default system prompt to the chat template to guide the model towards more professional, accurate, and safe responses.

    32.2BA8.8B128K ctx Apache 2.0
    25.6KSept 2025gguf
    • Same category
    • Similar size (32.2B vs 27.4B)
    • Commercial use OK
    vs
  20. FINAL-Bench logo

    Darwin-27B-RSI

    FINAL-Bench·Chat

    ≈18 GB

    Qwen3.5-27B family · 27B dense · Thinking mode · BF16 · Apache 2.0 No human-written answers. The model generated its own learning signal — and got measurably better.

    26.9B256K ctx Apache 2.0
    388Sept 2026gguf
    • Same category
    • Similar size (26.9B vs 27.4B)
    • Commercial use OK
    vs
  21. allenai logo

    Olmo-3-1125-32B

    allenai·Chat

    ≈22 GB

    We introduce Olmo 3, a new family of 7B and 32B models. This suite includes Base, Instruct, and Think variants. The Base models were trained using a staged training approach.

    32.2B64K ctx Apache 2.0
    16.6KNov 2025ollama
    • Same category
    • Similar size (32.2B vs 27.4B)
    • Commercial use OK
    vs
  22. google logo≈20 GB

    This section describes the MedGemma model and how to use it.

    27B128K ctx health-ai-developer-foundations
    14.9KMay 2025gguf
    • Same category
    • Similar size (27B vs 27.4B)
    vs
  23. ReadyArt logo≈27 GB

    scotoma‑1 writes very similarly to the base model. The J-Space projected refusal edit only lightly touched the prose.

    31.3B256K ctx Apache 2.0
    1.6KAug 2026gguf
    • Same category
    • Similar size (31.3B vs 27.4B)
    • Commercial use OK
    vs
  24. InternScience logo

    Agents-A1

    InternScience·Chat

    ≈22 GB

    Agents‑A1 is a 35B Mixture‑of‑Experts agentic model from InternScience, built to scale heterogeneous agentic abilities across multiple domains including Long‑horizon Search, Engineering, Scientific Research, Instruction Following, and Tool-calling.

    35.1BA3.9B256K ctx Apache 2.0
    5KJun 2026gguf
    • Same category
    • Similar size (35.1B vs 27.4B)
    • Commercial use OK
    vs