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 Qwen3.8-27B-DSpine-b16

120 open-weight models that do the same job as Qwen3.8-27B-DSpine-b16 (2B 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
tencent logo≈1.9 GB

This checkpoint is a five-layer draft model for Qwen3.8-27B, with block size 16 and an 8,192-token sliding window.

2B256K ctx Apache 2.0
4.9KNov 2026gguf
--vram
any
--gpus
--quant
  1. empero-ai logo

    Qwen3.8-4B-Distill

    empero-ai·Chat

    ≈4.3 GB

    Qwen3.8-4B is a full-parameter distillation of Qwen3.8 2.4T A95B into the Qwen3.5-4B architecture. The student was trained on ~45,000 curated teacher traces from our internal Qwen3.8 distillation datasets — dense chain-of-thought spanning mathematics, general reasoning, and instruction following…

    4.7B256K ctx Apache 2.0
    130.3KAug 2026gguf
    • Same category
    • Larger (4.7B vs 2B)
    • Commercial use OK
    vs
  2. HuggingFaceTB logo

    SmolLM2-1.7B

    HuggingFaceTB·Chat

    ≈3.0 GB

    The 1.7B variant demonstrates significant advances over its predecessor SmolLM1-1.7B, particularly in instruction following, knowledge, reasoning, and mathematics.

    1.7B8K ctx Apache 2.0
    42.6KOct 2024ollama
    • Same category
    • Similar size (1.7B vs 2B)
    • Commercial use OK
    vs
  3. LiquidAI logo

    LFM2-2.6B

    LiquidAI·Chat

    ≈2.5 GB

    LFM2 is a new generation of hybrid models developed by Liquid AI, specifically designed for edge AI and on-device deployment. It sets a new standard in terms of quality, speed, and memory efficiency.

    2.6B125K ctx lfm1.0
    6.7KSept 2025gguf
    • Same category
    • Similar size (2.6B vs 2B)
    vs
  4. meta-llama logo

    Llama-3.2-3B-Instruct

    meta-llama·Chat

    ≈3.3 GB

    The Llama 3.2 collection of multilingual large language models (LLMs) is a collection of pretrained and instruction-tuned generative models in 1B and 3B sizes (text in/text out).

    3.2B128K ctx Llama 3.2 Community License
    846.5KSept 2024ollama
    • Same category
    • Similar size (3.2B vs 2B)
    vs
  5. nvidia logo≈3.6 GB

    Nemotron-Labs-Diffusion is a tri-mode language model that supports both AR decoding and diffusion-based parallel decoding by simply switching the attention pattern of the same model during inference.

    3.8B256K ctx nvidia-nemotron-open-model-license
    70.4KMar 2026gguf
    • Same category
    • Similar size (3.8B vs 2B)
    • Commercial use OK
    vs
  6. Qwen logo≈1.5 GB

    In August 2024, we released the first series of mathematical LLMs - Qwen2-Math - of our Qwen family. A month later, we have upgraded it and open-sourced Qwen2.5-Math series, including base models Qwen2.5-Math-1.5B/7B/72B, instruction-tuned models Qwen2.5-Math-1.5B/7B/72B-Instruct, and mathematical…

    1.5B4K ctx Apache 2.0
    159.7KSept 2024gguf
    • Same category
    • Similar size (1.5B vs 2B)
    • Commercial use OK
    vs
  7. Qwen logo

    Qwen3-0.6B

    Qwen·Chat

    ≈1.8 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.

    752M40K ctx Apache 2.0
    18.5MApr 2025ollama
    • Same category
    • Smaller (752M vs 2B)
    • Commercial use OK
    • Widely used
    vs
  8. ibm-granite logo

    granite-swash-3b-a600m

    ibm-granite·Chat

    ≈2.5 GB

    Model Summary Granite-SWASH-3B-a600M is a 3B parameter decoder-only sparse language model with per-layer Sliding Window Attention (SWA) and learnable attention sinks (LSE-based).

    3BA600M8K ctx Apache 2.0
    3.1KJul 2026gguf
    • Same category
    • Similar size (3B vs 2B)
    • Commercial use OK
    vs
  9. Zyphra logo

    ZR1-1.5B

    Zyphra·Chat

    ≈1.8 GB

    ZR1-1.5B is a small reasoning model trained extensively on both verified coding and mathematics problems with reinforcement learning.

    1.8B128K ctx MIT
    8.3KApr 2025gguf
    • Same category
    • Similar size (1.8B vs 2B)
    • Commercial use OK
    vs
  10. Qwen logo≈3.1 GB

    Qwen1.5 is the beta version of Qwen2, a transformer-based decoder-only language model pretrained on a large amount of data. In comparison with the previous released Qwen, the improvements include:

    1.8B32K ctx Qwen Research License
    150.4KJan 2024gguf
    • Same category
    • Similar size (1.8B vs 2B)
    vs
  11. NVIDIA-Nemotron-Labs-Dory1.0-4B-BF16 is a looped (latent-recurrent) large language model (LLM) trained by NVIDIA. The model employs a normalized-GPT architecture with interleaved hybrid sliding window attention (SWA) and full attention layers.

    4B256K ctx openmdw-1.1
    15.6KOct 2026gguf
    • Same category
    • Similar size (4B vs 2B)
    • Commercial use OK
    vs
  12. google logo

    gemma-2-2b-it

    google·Chat

    ≈2.9 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.

    2.6B8K ctx Gemma Terms of Use
    170.2KJul 2024ollama
    • Same category
    • Similar size (2.6B vs 2B)
    vs
  13. IFM logo≈1.6 GB

    K2-Horizon-0.9B is the compact dense member of the K2-Horizon family: a 0.9B-class decoder-only model with a 128K context window.

    1.1B128K ctx Apache 2.0
    11.7KSept 2026gguf
    • Same category
    • Similar size (1.1B vs 2B)
    • Commercial use OK
    vs
  14. LiquidAI logo

    LFM2-2.6B-Exp

    LiquidAI·Chat

    ≈2.5 GB

    LFM2-2.6B-Exp is an experimental checkpoint built on LFM2-2.6B using pure reinforcement learning.

    2.6B125K ctx lfm1.0
    2.3KDec 2025gguf
    • Same category
    • Similar size (2.6B vs 2B)
    vs
  15. openbmb logo

    BitCPM-CANN-3B

    openbmb·Chat

    ≈2.5 GB

    BitCPM-CANN is the first end-to-end 1.58-bit (ternary) large language model training system natively built on Huawei Ascend NPU.

    3B32K ctx Apache 2.0
    3KMay 2026gguf
    • Same category
    • Similar size (3B vs 2B)
    • Commercial use OK
    vs
  16. Qwen logo≈1.5 GB

    In August 2024, we released the first series of mathematical LLMs - Qwen2-Math - of our Qwen family. A month later, we have upgraded it and open-sourced Qwen2.5-Math series, including base models Qwen2.5-Math-1.5B/7B/72B, instruction-tuned models Qwen2.5-Math-1.5B/7B/72B-Instruct, and mathematical…

    1.5B4K ctx Apache 2.0
    99.3KSept 2024gguf
    • Same category
    • Similar size (1.5B vs 2B)
    • Commercial use OK
    vs
  17. Salesforce logo

    CoDA-v0-Instruct

    Salesforce·Chat

    ≈2.6 GB

    Welcome to CoDA, Salesforce AI Research's diffusion-based language model designed for powerful code generation and bidirectional context understanding, presented in the paper CoDA: Coding LM via Diffusion Adaptation.

    2B40K ctx CC BY-NC 4.0
    501Sept 2025
    • Same category
    • Similar size (2B vs 2B)
    vs
  18. tencent logo≈2.1 GB

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

    1.8B256K ctx Unknown licence
    2.4KJul 2025gguf
    • Same category
    • Similar size (1.8B vs 2B)
    vs
  19. llm-jp logo≈2.4 GB

    LLM-jp-3.1 is a series of large language models developed by the Research and Development Center for Large Language Models at the National Institute of Informatics.

    1.9B4K ctx Apache 2.0
    2.5KMay 2025gguf
    • Same category
    • Similar size (1.9B vs 2B)
    • Commercial use OK
    vs
  20. WeiboAI logo

    VibeThinker-3B

    WeiboAI·Chat

    ≈2.6 GB

    2.For harder math reasoning, try AMOBench , a problem set harder than the International Mathematical Olympiad (IMO), with included standard answers. Use it to evaluate VibeThinker against other SOTA models. Note: due to extreme difficulty, set max tokens to 60K–100K.

    3.1B128K ctx MIT
    2.8KJun 2026gguf
    • Same category
    • Similar size (3.1B vs 2B)
    • Commercial use OK
    vs
  21. ibm-granite logo

    granite-4.1-3b-base

    ibm-granite·Chat

    ≈3.1 GB

    Model Summary: Granite‑4.1‑3B‑Base is a decoder‑only language model with long‑context capabilities, designed to support a broad range of general text‑to‑text generation tasks, as well as fill‑in‑the‑Middle (FIM) code completion.

    3.4B128K ctx Apache 2.0
    10.9KApr 2026gguf
    • Same category
    • Similar size (3.4B vs 2B)
    • Commercial use OK
    vs
  22. meta-llama logo

    Llama-3.2-1B

    meta-llama·Chat

    ≈1.5 GB

    The Llama 3.2 collection of multilingual large language models (LLMs) is a collection of pretrained and instruction-tuned generative models in 1B and 3B sizes (text in/text out).

    1.2B128K ctx Llama 3.2 Community License
    656.9KSept 2024gguf
    • Same category
    • Similar size (1.2B vs 2B)
    vs
  23. ai21labs logo

    AI21-Jamba2-3B

    ai21labs·Chat

    ≈2.4 GB

    Jamba2 3B is an ultra-compact open source model designed to bring enterprise-grade reliability to on-device deployments. At just 3B parameters, it runs efficiently on consumer devices—iPhones, Androids, Macs, and PCs—while maintaining the grounding and instruction-following capabilities required…

    3B256K ctx Apache 2.0
    6.8KJan 2026gguf
    • Same category
    • Similar size (3B vs 2B)
    • Commercial use OK
    vs
  24. kyutai logo

    helium-1-2b

    kyutai·Chat

    ≈2.1 GB

    Helium-1 is a lightweight language model with 2B parameters, targeting edge and mobile devices. It supports the 24 official languages of the European Union.

    2B4K ctx CC BY-SA 4.0
    1.2KApr 2025gguf
    • Same category
    • Similar size (2B vs 2B)
    • Commercial use OK
    vs