# Text Generation Models, Page 22: Open-Weight AI Source: https://savrn.com/models/tasks/text-generation/page/22 Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- SAVRN Model Hub · Models by Task # Text Generation Models 1,548 open-weight text generation models in the SAVRN Model Hub, with Convergent Intelligence, Qwen and Francesca Padovani publishing the most. 1,548Models 721Publishers 107K to 1.7TParameter range 8Licenses ## SAVRN's Take Text generation is where most of the attention goes, and where most of the money is spent. The models on this page run from tiny research checkpoints that fit on a laptop to frontier-class systems that need a full server of accelerators. The first question we ask about any of them is not how smart it is, but what it costs to keep it answering around the clock. That cost comes down to three things: the size of the weights, the precision you run them at, and how much context each request carries. Moving from 16-bit to 4-bit cuts the memory a model needs by roughly four times, which is often the difference between one server and several. Mixture-of-experts models, such as the newer DeepSeek, Qwen and Nemotron releases, hold a large total parameter count but activate only a small share per token, so they answer faster than their size suggests while still needing the memory to hold every expert. Licensing decides the rest. Apache 2.0 and MIT models can go straight into a commercial product. Llama, Gemma's earlier generations and models marked other each carry terms you need to read first, and a model with no stated license is research material until the publisher says otherwise. Our advice is to shortlist by license, size to the hardware you can actually run, then test the survivors on your own prompts. The table below is sorted by what people download, not by what we recommend. SAVRN Research, 2026-09-18 ## Most Downloaded | Model | Publisher | Parameters | License | Monthly downloads | Cheapest GPUs at 16-bit | | --- | --- | --- | --- | --- | --- | | [Qwen3-0.6B](https://savrn.com/models/qwen3-0-6b) | Qwen | 752M | apache-2.0 | 26.7M | 1x MI300X, $1.85/hr | | [gpt2](https://savrn.com/models/gpt2) | OpenAI community | 137M | mit | 18.6M | 1x MI300X, $1.85/hr | | [Qwen3-4B](https://savrn.com/models/qwen3-4b) | Qwen | 4B | apache-2.0 | 11.6M | 1x MI300X, $1.85/hr | | [Qwen3-14B](https://savrn.com/models/qwen3-14b) | Qwen | 14.8B | apache-2.0 | 9.2M | 1x MI300X, $1.85/hr | | [Qwen3-8B](https://savrn.com/models/qwen3-8b) | Qwen | 8.2B | apache-2.0 | 8.3M | 1x MI300X, $1.85/hr | | [Qwen2.5-7B-Instruct](https://savrn.com/models/qwen2-5-7b-instruct) | Qwen | 7.6B | apache-2.0 | 6.6M | 1x MI300X, $1.85/hr | | [gpt-oss-20b](https://savrn.com/models/gpt-oss-20b) | OpenAI | 20.9B | apache-2.0 | 5.8M | 1x MI300X, $1.85/hr | | [opt-125m](https://savrn.com/models/opt-125m) | AI at Meta | — | other | 5.8M | — | | [Qwen2.5-0.5B-Instruct](https://savrn.com/models/qwen2-5-0-5b-instruct) | Qwen | 494M | apache-2.0 | 5.5M | 1x MI300X, $1.85/hr | | [DeepSeek-V3.2](https://savrn.com/models/deepseek-v3-2) | DeepSeek | 685.4B | mit | 5.5M | 7x MI325X, $14.00/hr | ## Licenses | License | Models | Commercial use | | --- | --- | --- | | [apache-2.0](https://savrn.com/models/licenses/apache-2-0) | 714 | Yes | | not stated | 341 | Not stated | | other | 168 | Read the license | | [mit](https://savrn.com/models/licenses/mit) | 151 | Yes | | [cc-by-nc-4.0](https://savrn.com/models/licenses/cc-by-nc-4-0) | 41 | Not without separate permission | | [llama3.2](https://savrn.com/models/licenses/llama3-2) | 33 | Read the license | | [gemma](https://savrn.com/models/licenses/gemma) | 22 | Yes, with conditions | | [llama3.1](https://savrn.com/models/licenses/llama3-1) | 21 | Yes, with conditions | ## Who Publishes Them | Publisher | Models | | --- | --- | | [Convergent Intelligence](https://savrn.com/model-publishers/reaperdoesntknow) | 54 | | [Qwen](https://savrn.com/model-publishers/qwen) | 42 | | [Francesca Padovani](https://savrn.com/model-publishers/francesca9805) | 37 | | [Yanggangu](https://savrn.com/model-publishers/yanggangu) | 28 | | [Shreyansh singh](https://savrn.com/model-publishers/shreyansh12183) | 25 | | [Kumar](https://savrn.com/model-publishers/sandeep123) | 23 | ## All 1,548 Models, Page 22 of 26 Model · Text generation ### [qwen3.6-35b-a3b-tool-prompts-adapter](https://savrn.com/models/qwen3-6-35b-a3b-tool-prompts-adapter) [Manish](https://savrn.com/model-publishers/manishkumar2101114) Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Use the code below to get started with the model. Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - PEFT 0.20.0 Open weights peft [View model](https://savrn.com/models/qwen3-6-35b-a3b-tool-prompts-adapter) Model · Text generation ### [qwen3.6-35b-a3b-tool-prompts-ckpt444-adapter](https://savrn.com/models/qwen3-6-35b-a3b-tool-prompts-ckpt444-adapter) [Manish](https://savrn.com/model-publishers/manishkumar2101114) Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Use the code below to get started with the model. Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - PEFT 0.20.0 Open weights peft [View model](https://savrn.com/models/qwen3-6-35b-a3b-tool-prompts-ckpt444-adapter) Model · Text generation ### [Qwen3.6-35B-A3B-TT-Mixed-BFP4-BFP8-P150](https://savrn.com/models/qwen3-6-35b-a3b-tt-mixed-bfp4-bfp8-p150) [Lotto](https://savrn.com/model-publishers/lottolabs) A hardware-specific, text-only checkpoint and custom runtime derived from Qwen/Qwen3.6-35B-A3B, revision 995ad96eacd98c81ed38be0c5b274b04031597b0, for one Tenstorrent Blackhole P150. This is a community experimental release by Lottolabs, not an official Qwen or Tenstorrent release. It is not GGUF, AWQ, an AutoGPTQ/GPTQModel int4 checkpoint, bitsandbytes, NVFP4, or an ordinary Transformers checkpoint. transformers.frompretrained() and hosted Inference Providers are not supported launch paths. The vision encoder of the upstream model is not included: the server accepts text only. - Checkpoint (checkpoint/, 21.7 GB, manifest sha256 439096088c8b…): - routed experts (256 per layer × 40 layers)… Open weights apache-2.0 [View model](https://savrn.com/models/qwen3-6-35b-a3b-tt-mixed-bfp4-bfp8-p150) Model · Text generation ### [Qwen3.8-27B-24GB-128k-GGUF](https://savrn.com/models/qwen3-8-27b-24gb-128k-gguf) [Jordan Dominick](https://savrn.com/model-publishers/jdominick05) Three GGUF quantizations of Qwen/Qwen3.8-27B that spend bits where quantization error hurts most: the 16 full-attention layers and the recurrent (linear-attention) layers get more precision, the bulky feed-forward up/gate weights get less. Built for running the model at a 128k context entirely on one 24 GB GPU (measured on an RX 7900 XTX with llama.cpp/ROCm). All are text-only (no vision projector). Unsloth's mmproj-F16.gguf from unsloth/Qwen3.8-27B-GGUF is built from the same base model and should work with them, but this was not tested. R1 and R2 keep the multi-token-prediction (MTP) layer, quantized with the same recipe; R3 drops it (64 blocks, nextnpredictlayers = 0), so it cannot be… Open weights apache-2.0 gguf [View model](https://savrn.com/models/qwen3-8-27b-24gb-128k-gguf) Model · Text generation ### [Qwen3.8-27B-depthaware-GGUF](https://savrn.com/models/qwen3-8-27b-depthaware-gguf) [Federico Sciuca](https://savrn.com/model-publishers/federicosciuca) A 27B dense model compressed to 3.66 effective bits, built from a measurement rather than a convention - and the build that answered a question this project had been avoiding. Most low-bit quants decide which layers to protect by rule of thumb. This one was built from a probe: every depth band of this specific model was pushed to 2-bit one at a time and scored on held-out text, and the protection went where the damage actually was (the last quarter of the network, blocks 51-63). Every equal-bytes win behind this method had been measured on mixture-of-experts models. MoE has a structural feature a dense model lacks - most of each layer's weight sits in routed experts that move between memory… Open weights gguf [View model](https://savrn.com/models/qwen3-8-27b-depthaware-gguf) Model · Text generation ### [Qwen3.8-27B-Flash-MTP-GGUF](https://savrn.com/models/qwen3-8-27b-flash-mtp-gguf) [SuperLogic](https://savrn.com/model-publishers/superlogic) 让普通用户进入 0 Token 时代。 不需要租 API、不需要云、不需要买卡。一台 8 GB 显存的普通笔记本，在自己家里跑 27B 级推理模型： 作为对照：同一台机器上，LM Studio 跑 Qwen3.8-27B 的 4-bit 量化，大约 1–2 t/s。 最后一条是关键：MTP 草稿头的 output.weight 是全词表 [5120, 248320]（265 MB），每个起草步都要整读一遍。 目标模型验证时仍然用全词表，所以输出分布一个 bit 都不变 —— 只省字节，不伤质量（实测接受率持平）。 上图：开 low 思考档，模型生成的中国象棋棋盘布局。 脚本和二进制在 GitHub —— 刻意分成两个仓，因为脚本得能被人读、被人改： features.ini 不在引擎源码仓里，也不是编译进二进制的。 引擎没有它的内置默认路径， 它是启动时用 --features-file features.ini 读进来的。少了它，自己编出来的二进制跑的是 上游行为，不是我们的调校（比如 spec-draft-vocab-map 那条 24k 起草短名单、kv-vram-cells 的 KV 分箱、spec-draft-sampling = probabilistic）。 复现要点（细节见 runtime 仓的 README）： - 采样 --temp 0.6 --top-k 20 --top-p 0.95 - KV 缓存用 q40，长上下文靠 tiered KV（kv-vram-cells）把超出的部分放 pinned host 内存 - 自投机必须带… Open weights apache-2.0 llama.cpp [View model](https://savrn.com/models/qwen3-8-27b-flash-mtp-gguf) Model · Text generation ### [qwen3.8-27b-novel-lora](https://savrn.com/models/qwen3-8-27b-novel-lora) [Hui](https://savrn.com/model-publishers/wodeapp) 把 Qwen3.8-27B 转为小说训练后的 LoRA 适配器，用于中文小说正文写作。 - 基座：Qwen3.8-27B - 类型：LoRA（PEFT） - alpha：128 加载时把适配器挂到对应的 Qwen3.8-27B 基座上。本仓库只有适配器，不含基座权重。 Open weights other peft [View model](https://savrn.com/models/qwen3-8-27b-novel-lora) Model · Text generation ### [Qwen3.8-27B-OBLITERATED](https://savrn.com/models/qwen3-8-27b-obliterated-2) [Ahmed](https://savrn.com/model-publishers/ahmed-amedo) V3 applies iterative refinement on top of V2's complementary blend, with targeted corpus expansion. The result: genuine liberation — not just removal of hard refusals but elimination of safety-lecture deflections. - Genuinely answers restricted queries — provides real substance instead of safety lectures - 20/20 on code generation tasks — functional implementations, not disclaimers - Thinking ON compatible — no refusals in either thinking mode - Honest scoring — every response manually audited for real substance, not just absence of "I cannot" - -2.1pp MMLU — modest capability cost for genuine liberation If you're using this model in an agent harness (coding agent, pentest framework, etc.)… Open weights apache-2.0 27.8B parameters 262,144 tokens mlx [View model](https://savrn.com/models/qwen3-8-27b-obliterated-2) Model · Text generation ### [qwen3.8-27b-pi-p300](https://savrn.com/models/qwen3-8-27b-pi-p300) [Taylor Singletary](https://savrn.com/model-publishers/episod) Open weights apache-2.0 [View model](https://savrn.com/models/qwen3-8-27b-pi-p300) Model · Text generation ### [Qwen3.8-27B-Ternary-Bonsai-2-DFlash2-MLX](https://savrn.com/models/qwen3-8-27b-ternary-bonsai-2-dflash2-mlx) [Nate Sutton](https://savrn.com/model-publishers/nathansutton) Prism ML's ternary Ternary-Bonsai-2-27B build of Qwen/Qwen3.8-27B, repacked for chad, a Claude-Code-style local coding agent for Apple Silicon, with its speculative decoder bundled in. This is chad's default model. Created using Bonsai by Prism ML. with, already quantized. Nothing is built on first run. Every projection of Qwen3.8-27B (a dense qwen35 hybrid: 64 layers, 48 GatedDeltaNet + 16 full attention) is stored in a Hadamard-rotated basis: multiplied by a fixed sign vector and put through a blockwise Walsh-Hadamard transform offline, then quantized to 2-bit affine group-128 whose three levels reproduce the ternary set {−s, 0, +s}. The rotation costs no extra bits and no extra weight… Open weights apache-2.0 26.9B parameters 262,144 tokens mlx [View model](https://savrn.com/models/qwen3-8-27b-ternary-bonsai-2-dflash2-mlx) Model · Text generation ### [Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NEO-CODER-MAX-MTP-MLX](https://savrn.com/models/qwen3-8-27b-turbo-fable-cold-fusion-735-882-heretic-uncensored-neo-coder-max-mtp-mlx) [Elvis McFee](https://savrn.com/model-publishers/ecmcfee) MLX-optimized version of DavidAU/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NEO-CODER-MAX-MTP for Apple Silicon inference. Converted from the Q80 GGUF quantization using gguf2mlx. All 866 tensors are dequantized to float16, so this model has identical quality to the original GGUF weights — no precision loss from the conversion. Apache-2.0 (same as the original model) Open weights apache-2.0 27.3B parameters 262,144 tokens mlx [View model](https://savrn.com/models/qwen3-8-27b-turbo-fable-cold-fusion-735-882-heretic-uncensored-neo-coder-max-mtp-mlx) Model · Text generation ### [Qwen3.8-27B-UD-Q8_K_XL-layers](https://savrn.com/models/qwen3-8-27b-ud-q8-k-xl-layers) [Maveric Studio](https://savrn.com/model-publishers/studiomaveric) GGUF layer package for running Qwen3.8-27B-UD-Q8KXL across a local Mesh LLM cluster. This package is derived from unsloth/Qwen3.8-27B-GGUF and keeps the original GGUF distribution split into per-layer artifacts for distributed inference. - Local and private inference with Mesh LLM. - Multi-machine serving when the full GGUF is too large for one host. - OpenAI-compatible chat/completions workflows through Mesh LLM's local API. For upstream architecture details, chat template guidance, sampling recommendations, license terms, and benchmark notes, see the source model card: unsloth/Qwen3.8-27B-GGUF. Generated by the Mesh LLM HF Jobs splitter from mesh-llm ref main. Each artifact is checksummed… Open weights apache-2.0 mesh-llm [View model](https://savrn.com/models/qwen3-8-27b-ud-q8-k-xl-layers) Model · Text generation ### [Qwen3.8-27B-Uncensored-Chinese](https://savrn.com/models/qwen3-8-27b-uncensored-chinese) [Vurtne Saerdna](https://savrn.com/model-publishers/vurtnesaerdna) 18+ only. This model generates sexually explicit fiction. A fine-tune of Qwen/Qwen3.8-27B for Chinese adult creative writing. It was trained with LoRA on about 4.2k instruction examples built from Chinese adult fiction, and the LoRA was then merged into the base weights, so this repo loads like a normal full model. Requires a recent transformers with Qwen3.8 support. The ~50 GB of bf16 weights need multiple GPUs or CPU offload. Avoid greedy decoding, which makes Qwen models more likely to repeat themselves. - Content is fictional and intended for adult readers only. - Trained mostly on long-form prose; instruction following outside creative writing may be weaker than the base model. - May… Open weights apache-2.0 27.8B parameters 262,144 tokens transformers [View model](https://savrn.com/models/qwen3-8-27b-uncensored-chinese) Model · Text generation ### [Qwen3.8-27B-Uncensored-GGUF](https://savrn.com/models/qwen3-8-27b-uncensored-gguf-2) [Tạ Hoà Bình](https://savrn.com/model-publishers/bnhho) Uncensored Qwen3.8-27B, published as GGUF quantizations with the multi token prediction (MTP) head retained and verified. Refusal behaviour has been substantially reduced, not eliminated. See Measured behaviour for the numbers. Capabilities, training data, and architecture are otherwise unchanged. - Refusal directions removed with Heretic, which co minimizes refusal count against KL divergence from the base model. No handwritten refusal removal code, no finetuning, no additional training data. - Abliteration runs at bf16 (no 4 bit quantization). the resulting LoRA is merged into the bf16 base, so the published weights are not a quantized round trip. - mtp. tensors are copied verbatim from… Open weights apache-2.0 llama.cpp [View model](https://savrn.com/models/qwen3-8-27b-uncensored-gguf-2) Model · Text generation ### [Qwen3.8-27B-VAL-APEX-I-MiniPlus-V3-GGUF](https://savrn.com/models/qwen3-8-27b-val-apex-i-miniplus-v3-gguf) [IsValorum](https://savrn.com/model-publishers/isvalorum) Official VAL-APEX-I MiniPlus V3 quantization of Qwen/Qwen3.8-27B. This edition quantizes the original Qwen checkpoint without applying Huihui abliteration or EfficientThink fine-tuning. Vector-calibrated Asymmetric Layer-wise Outlier-preserving Recurrent-aware Unified Matrix-quantization The NanoPlus V3 file is 11.35 GB, approximately 0.45 GB smaller than GSQ-RCO IQ3S, with 11.49% lower reported WikiText-2 PPL (6.2577 vs 7.07). MiniPlus V3 reaches 13.21% lower reported PPL (6.1361 vs 7.07) at 14.16 GB, approximately 2.36 GB larger. × is the number of tensors in each format. Counts were read from the released GGUF headers; the GSQ-RCO column uses its published tensor allocation. Both V3… Open weights apache-2.0 gguf [View model](https://savrn.com/models/qwen3-8-27b-val-apex-i-miniplus-v3-gguf) Model · Text generation ### [Qwen3.8-27B-VAL-APEX-I-NanoPlus-V3-GGUF](https://savrn.com/models/qwen3-8-27b-val-apex-i-nanoplus-v3-gguf) [IsValorum](https://savrn.com/model-publishers/isvalorum) Official VAL-APEX-I NanoPlus V3 quantization of Qwen/Qwen3.8-27B. This edition quantizes the original Qwen checkpoint without applying Huihui abliteration or EfficientThink fine-tuning. Vector-calibrated Asymmetric Layer-wise Outlier-preserving Recurrent-aware Unified Matrix-quantization The NanoPlus V3 file is 11.35 GB, approximately 0.45 GB smaller than GSQ-RCO IQ3S, with 11.49% lower reported WikiText-2 PPL (6.2577 vs 7.07). MiniPlus V3 reaches 13.21% lower reported PPL (6.1361 vs 7.07) at 14.16 GB, approximately 2.36 GB larger. × is the number of tensors in each format. Counts were read from the released GGUF headers; the GSQ-RCO column uses its published tensor allocation. Both V3… Open weights apache-2.0 gguf [View model](https://savrn.com/models/qwen3-8-27b-val-apex-i-nanoplus-v3-gguf) Model · Text generation ### [qwen3.8-flash-next-40b-prune-research](https://savrn.com/models/qwen3-8-flash-next-40b-prune-research) [DavidB](https://savrn.com/model-publishers/davd-b01) One upstream MoE, six generations of masks, one bit-exact surgery, a rebuilt factual memory, and a three-stage post-training (SFT → SimPO → RLVR) — with every error I made and every idea I falsified along the way. This repository is the complete technical record of the project: the activation-instrumentation artifacts, the mask-generation methodology, the surgery code, the PLE (n-gram memory) rebuild, the trainers, and the full lab notebook. The champion weights live in sibling repos What I set out to do: take Qwen3.8-Flash-Next — an ~180B-class MoE (~130B of core network + 51B of n-gram memory table + 4B MTP) — and cut it by 75% of its experts (512 → 128) while conserving the maximum… Open weights apache-2.0 40.7B parameters 262,144 tokens transformers [View model](https://savrn.com/models/qwen3-8-flash-next-40b-prune-research) Model · Text generation ### [Qwen3.8-Flash-Next-Abliterated-EXL3-2.50bpw](https://savrn.com/models/qwen3-8-flash-next-abliterated-exl3-2-50bpw) [Ghost](https://savrn.com/model-publishers/ghost-actual) A refusal-removed (abliterated) build of Qwen's Qwen3.8-Flash-Next, quantized to EXL3 2.50 bpw so the full model runs on a single 24 GB card (RTX 3090 / 4090) using MoE CPU-offload — with the vision tower, MTP head, native 262,144-token context, and the PLE n-gram table all intact. Requires the same MoE CPU-offload setup as the stock 2.50bpw pack. Needs ~59 GB host RAM for the CPU expert tail and a fast NVMe for the streamed n-gram table. Expected on an RTX 3090: ~38 tok/s decode with MTP on (~28 without), ~20 tok/s at 175K depth, ~664 tok/s prefill. See the upstream repo for the full measured ledger; this quant uses the identical flags and layout, so numbers should track closely. Fired on… Open weights other 22.3B parameters 262,144 tokens [View model](https://savrn.com/models/qwen3-8-flash-next-abliterated-exl3-2-50bpw) Model · Text generation ### [Qwen3.8-Flash-Next-CNQ4.5-M](https://savrn.com/models/qwen3-8-flash-next-cnq4-5-m) [Robin](https://savrn.com/model-publishers/nibor1896) Qwen3.8-Flash-Next quantized to CNQ4.5-M: one NVFP4 container at 4.5 bpw with a BF16 keep set and MSE sub-block scales (--scales mse), weights for the crow-nest engine on Linux and Windows. - Verify the download with the first command below, in the package directory; it must report 6 of 6 OK and covers the container, the sidecar, the hot-set manifest and the three self-test files. The second command checks the self-test files alone, which costs no read of the 105 GB container. - The container is the CNQ v1 format written by the converter in the crow-nest repository (converter/src/main.rs, module comment of record main.rs:1-59). That converter reproduces this package from the original… Open weights other [View model](https://savrn.com/models/qwen3-8-flash-next-cnq4-5-m) Model · Text generation ### [Qwen3.8-Flash-Next-Coder-P150](https://savrn.com/models/qwen3-8-flash-next-coder-p150) [Lotto](https://savrn.com/model-publishers/lottolabs) Block-float weights that let Strata's tt backend run Qwen3.8-Flash-Next Coder on one Tenstorrent Blackhole P150 card: - the routed experts of ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-Coder-GGUF (IQ1M), 48 layers x 256 experts x gate/up/down, re-quantized with GPTQ to Tenstorrent's BFP4 (bfp4b) format, in the exact byte layout the P150 kernels read; - the precision/placement plan bfp4-tiered-2.4e+10: which experts live in the card's DRAM and which in pinned host memory; - the MTP draft layer's 512 routed experts in BFP4 and their routing profile (speculative decoding). This is not a standalone model. It holds no tokenizer, attention, embedding or LM-head weights: those come from the GGUF that… Open weights other [View model](https://savrn.com/models/qwen3-8-flash-next-coder-p150) Model · Text generation ### [Qwen3.8-Flash-Next-DFlash2-Infernix](https://savrn.com/models/qwen3-8-flash-next-dflash2-infernix) [Ian](https://savrn.com/model-publishers/wallawalla47) Open weights apache-2.0 infernix [View model](https://savrn.com/models/qwen3-8-flash-next-dflash2-infernix) Model · Text generation ### [Qwen3.8-Flash-Next-DGX-Spark-Optimized-Recipe](https://savrn.com/models/qwen3-8-flash-next-dgx-spark-optimized-recipe) [Single Spark AI](https://savrn.com/model-publishers/single-spark-ai) A personal, measured-on-one-machine deployment recipe for serving Qwen/Qwen3.8-Flash-Next on a single NVIDIA DGX Spark (GB10) with SGLang. It is not a benchmark leaderboard claim and not "the fastest possible"; it is what is verifiably running on one GB10, with every patch, script, and check needed to reproduce it from upstream weights. Verified: 2026-09-13 · SGLang 00143e9c23aee2dead5e6fe217bda4fa8739cb92 (nightly-dev-cu13-20260911) · runtime image qwen38-flashnext-sm121-hybrid-sharp:00143e9c-hc2, built by bounded-ple/nightly/01c-build-sglang-nightly.sh. The v4 lineage (d91c3682-hc1) is still in the repo as the stable fallback. Key runtime facts (all asserted at deploy time by the 05… Open weights other [View model](https://savrn.com/models/qwen3-8-flash-next-dgx-spark-optimized-recipe) Model · Text generation ### [Qwen3.8-Flash-Next-HC-Q8](https://savrn.com/models/qwen3-8-flash-next-hc-q8) [Charles Pasquale](https://savrn.com/model-publishers/cafonez) Quantized GGUF of Qwen/Qwen3.8-Flash-Next. This is not an official Qwen release. The weights remain under the Qwen Community License 1.0. The file is Qwen3.8-Flash-Next-HC-Q8.gguf. - 179.551 billion parameters - 92,347,778,272 bytes - 4.11 bits per weight on average - Architecture qwen4exp, context trained at 262,144 HC-Q8 starts from a Q40ROCMI4 quant. The 200 hyperconnection up and down matrices were restored from the original BF16 tensors in Qwen revision de4b8e4d and stored as ordinary Q80. The 98 injection tensors stayed BF16. Every other tensor kept the ROCmI4 bytes. Q40ROCMI4 and Q30ROCMFPX are ROCmFPX quant types. Stock llama.cpp does not load this file. Use the ROCmFPX v2 HIP… Open weights other [View model](https://savrn.com/models/qwen3-8-flash-next-hc-q8) Model · Text generation ### [Qwen3.8-Flash-Next-heretic2-W4A16-FP8PLE](https://savrn.com/models/qwen3-8-flash-next-heretic2-w4a16-fp8ple) [FL](https://savrn.com/model-publishers/densohax) Run Qwen 3.8 Flash Next heretic-2 locally on two RTX 3090 24 GB GPUs and 128 GB RAM. Start with the GitHub quickstart and pinned runtime; the weights require its custom vLLM overlay. This checkpoint is the heretic-2 reassembly of the (AutoRound W4A16) instead of Intel's stock AutoRound target. 1,860 tok/s prefill · 89.1 tok/s decode · 2× RTX 3090 + 128 GB RAM decode after 131,072 input tokens. Each generates 2,048 tokens. The latest experimental runtime reaches the first token in 139.8 seconds with 260,096 input tokens, then generates 2,048 tokens at 86.2 tok/s. That fills the native 262,144-token window. The full request takes 163.6 seconds. The wait fell from 214.5 to 139.8 seconds… Open weights other 262,144 tokens [View model](https://savrn.com/models/qwen3-8-flash-next-heretic2-w4a16-fp8ple) Model · Text generation ### [Qwen3.8-Flash-Next-ik-llama-GGUF_v2](https://savrn.com/models/qwen3-8-flash-next-ik-llama-gguf-v2) [Joel Farthing](https://savrn.com/model-publishers/ji-farthing) HuggingFace's download widget does not recognize KT GGUFs. Download links are in the table below. Text-only GGUF conversion of Qwen3.8-Flash-Next, quantized with an importance matrix calibrated for agentic coding. This version uses IQ4KT for the 160-column PLE table and IQ3KT for the expert-down and shared-expert-down tensors. The previous release used IQ4NL for tensors whose row widths were not multiples of 256. The new file is 80.72 GiB, compared with 88.42 GiB for the previous IQ4KT. This file requires the IQ3KT/IQ4KT 32-element row-tail changes. Runtime validation used candidate 0b7b979. The source branch and PR are pending publication. This upload is the test artifact for that… Open weights other [View model](https://savrn.com/models/qwen3-8-flash-next-ik-llama-gguf-v2) Model · Text generation ### [Qwen3.8-Flash-Next-MoEspressoV3](https://savrn.com/models/qwen3-8-flash-next-moespressov3) [Riccardo](https://savrn.com/model-publishers/steadfastgaze) This package brings Qwen3.8-Flash-Next to MoEspresso. Qwen3.8-Flash-Next is a 125B Mixture-of-Experts model with 6B active parameters per token and a 51B-parameter PLE/n-gram memory. On a 2021 M1 Max with a 24-core GPU and 32 GB of unified memory, MoEspresso serves this package at roughly 12–15 decode tokens/s. The local bounded configuration also scored 84.3% on a frozen 48-question generated-answer comparison, ahead of Claude Opus 4.8 xhigh at 80.3% on the same questions. The complete download occupies 145.75 GB (135.74 GiB), including the original BF16 PLE tables. MoEspresso keeps only its active working set in unified memory. The package requires MoEspresso 3.0.0 on an arm64 Apple… Open weights other 262,144 tokens mlx [View model](https://savrn.com/models/qwen3-8-flash-next-moespressov3) Model · Text generation ### [Qwen3.8-Flash-Next-NInfer-Offload](https://savrn.com/models/qwen3-8-flash-next-ninfer-offload) [Cha Luo](https://savrn.com/model-publishers/luocha2050) English | 中文 Qwen/Qwen3.8-Flash-Next converted to the.ninfer v3 format for NInfer-Offload, which runs this 125B-total / 6B-active MoE model on a 16 GB consumer GPU by keeping the routed experts in system RAM. It includes the model's MTP predictor for speculative decoding. This file only works with NInfer-Offload. It is not a GGUF or safetensors checkpoint and cannot be loaded by llama.cpp, vLLM, transformers or upstream NInfer. The 110 GB artifact is stored as NInfer's native multi-file set (at most 50 GB per file). Keep all files in one directory and pass only the entry file qwen38flashnext.ninfer to NInfer; it reads the.part-000N continuation files itself. No reassembly is needed.… Open weights other ninfer [View model](https://savrn.com/models/qwen3-8-flash-next-ninfer-offload) Model · Text generation ### [Qwen3.8-Flash-Next-Q2_0-H64-S4-GGUF](https://savrn.com/models/qwen3-8-flash-next-q2-0-h64-s4-gguf) [Facu Vlad J](https://savrn.com/model-publishers/groxaxo) I wanted a smaller GGUF I could run, and I wanted to know more than just whether it loaded. This is where that experiment landed. How we got here has the story, including the first Q2 export, the H64-S4 experiment, the interrupted run, and what I could and couldn't verify. You need a runtime that supports the matching Hadamard transform. A runtime that ignores the rotation metadata can compute the wrong result. The tested llama.cpp changes, source patch, and build instructions are in runtime/. The export has nine GGUF shards, about 68.9 GB in total. It is based on Qwen/Qwen3.8-Flash-Next, revision de4b8e4d43b917e7706784d8bb445c9af86a3540. I used mixed precision to spend the bits where the… Open weights other gguf [View model](https://savrn.com/models/qwen3-8-flash-next-q2-0-h64-s4-gguf) Model · Text generation ### [qwen3.8-flash-next-radiance-q5-w5dn64](https://savrn.com/models/qwen3-8-flash-next-radiance-q5-w5dn64) [Gostar](https://savrn.com/model-publishers/gostar77) A checkpoint-specific calibrated quantization of Qwen/Qwen3.8-Flash-Next for the Radiance inference engine. This repository contains a single self-contained.rad inference container. It is not a Transformers, safetensors, llama.cpp, or vLLM checkpoint. Use a Radiance build with the Qwen3.8-Flash-Next model support and w4nl/w5nl expert kernels. - All 48 calibrated trunk layers use GPTQ with three coordinate-descent sweeps. - Routed expert down projections use non-linear w5nl (5-bit) codebooks, group 64, E4M3 scales, and FWHT128. - Routed expert gate/up projections use w4nl (4-bit) with the same transform and group structure. - Ten selected high-energy routed experts remain BF16. - Attention… Open weights other radiance [View model](https://savrn.com/models/qwen3-8-flash-next-radiance-q5-w5dn64) Model · Text generation ### [Qwen3.8-Flash-Next-W8A8-DYNAMIC-300i](https://savrn.com/models/qwen3-8-flash-next-w8a8-dynamic-300i) [Matt Davis](https://savrn.com/model-publishers/matteiuspi) This repository contains an Ascend ModelSlim W8A8DYNAMIC conversion of pinned at revision It targets two Atlas 300I Duo cards, exposed as four Ascend 310P3 devices, with tensor parallelism across all four devices. The checkpoint preserves the base model's Qwen4 experimental architecture: 48 decoder layers, 512 routed experts, hybrid Gated DeltaNet/Qwen Sparse Attention, PLE n-gram embeddings, a vision encoder, and MTP weights. The current Ascend serving path has only been exercised for text generation. Multimodal execution is not claimed as supported, and MTP remains provisional pending the quality comparison and final public runtime freeze. The conversion used Ascend ModelSlim's… Open weights other 262,144 tokens vllm [View model](https://savrn.com/models/qwen3-8-flash-next-w8a8-dynamic-300i) Model · Text generation ### [qwen3-8b-sft-261001-bigmath-sol-fsdp-0-epoch-4](https://savrn.com/models/qwen3-8b-sft-261001-bigmath-sol-fsdp-0-epoch-4) [Seungyeop Yi](https://savrn.com/model-publishers/devpotatopotato) Qwen3-8B after full-parameter supervised fine-tuning for mathematical keyword and meaning generation. This is the exact saved epoch 4 checkpoint at optimizer step 2976. The repository contains the model weights, configuration, and tokenizer files for inference. Training dataset: devpotatopotato/math-keyword-training, keyword-261001-bigmath-gpt-6-sol.jsonl. Use the tokenizer chat template with enablethinking=False. Pass the following template as a user message and replace {problem} with the problem text Open weights apache-2.0 4.1B parameters 40,960 tokens transformers [View model](https://savrn.com/models/qwen3-8b-sft-261001-bigmath-sol-fsdp-0-epoch-4) Model · Text generation ### [Qwen3-8b-urban-seed1-20261010-nvl-koqf8weu-CPT-merged-epoch-2](https://savrn.com/models/qwen3-8b-urban-seed1-20261010-nvl-koqf8weu-cpt-merged-epoch-2) [Brandon Howe](https://savrn.com/model-publishers/brandonhowe) Standalone merged BF16 model from epoch 2.0, step 756. Dataset: CompassioninMachineLearning/urban12738cleaned at ef7c0e742df63ea319e35d02d9f6ba63d6e7c68d. Training: 10,072 distinct documents plus 2,000 repeat exposures per epoch; 200 disjoint validation documents. Merged with Unsloth's native savepretrainedmerged(savemethod="merged16bit"). Weights are validated BF16 and packaged losslessly into eight safetensors shards. No adapter is required to load this model. See runmanifest.json for base revision, document selection hashes, training parameters and export validation. Training does not establish an improvement in compassion; evaluate that separately. Open weights 8.2B parameters 32,768 tokens transformers [View model](https://savrn.com/models/qwen3-8b-urban-seed1-20261010-nvl-koqf8weu-cpt-merged-epoch-2) Model · Text generation ### [qwen3-8b-worker-farca](https://savrn.com/models/qwen3-8b-worker-farca) [Atlas Institute of Mechatronics](https://savrn.com/model-publishers/atlas-institute) Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Use the code below to get started with the model. Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - PEFT 0.19.1 Open weights peft [View model](https://savrn.com/models/qwen3-8b-worker-farca) Model · Text generation ### [Qwen3-Coder-30B-A3B-Instruct-abliterated-GGUF](https://savrn.com/models/qwen3-coder-30b-a3b-instruct-abliterated-gguf) [Alan Frederick Benitez](https://savrn.com/model-publishers/thaurock) Esta es la colección integral y completa de cuantizaciones en formato GGUF del modelo Huihui-Qwen3-Coder-30B-A3B-Instruct-abliterated, preparadas localmente para su uso con llama.cpp, Ollama, LM Studio o Text-Generation-WebUI. Este modelo combina las avanzadas capacidades de programación y razonamiento lógico de la arquitectura Qwen3-Coder (MoE), pero procesada con técnicas de abliteration para remover de raíz los bloqueos y filtros de censura artificiales del sistema, respondiendo sin restricciones. Nota: Los tamaños son estimaciones iniciales de referencia basadas en el peso nativo del modelo en safetensors (31B parámetros MoE); se recomienda verificar el tamaño final en disco tras la… Open weights apache-2.0 [View model](https://savrn.com/models/qwen3-coder-30b-a3b-instruct-abliterated-gguf) Model · Text generation ### [Qwen3-Coder-Next-JNS](https://savrn.com/models/qwen3-coder-next-jns) [Prabanta Dev](https://savrn.com/model-publishers/prabanta-dev) On a laptop running Debian 13 Linux, with an Intel Core Ultra 9 185H, its integrated Arc GPU and 32 GB of memory, Janas-LLM reads a prompt at 66 tok/s and writes at 22 tok/s: llama.cpp, on the same laptop, 14 and 2.2. Measured on 9-11 October 2026: llama-bench -p 1024 -n 16 and janas-bench --prompt 1024 --gen 16, both at their best configuration. On other machines the figures will differ. Details: Janas README. - Debian 13 (trixie) Linux, x86-64, kernel 6.12.107 - its integrated Intel Arc GPU, through the Vulkan driver of Mesa 26.1.6 - 32 GB of DDR5-5600 memory - a 1 TB NVMe SSD, about 5.7 GB/s reading - llama.cpp build 2b18470 (18 September 2026), Vulkan backend The model (48.4 GB) is… Open weights apache-2.0 [View model](https://savrn.com/models/qwen3-coder-next-jns) Model · Text generation ### [Qwen3-Next-80B-A3B-Instruct-JNS](https://savrn.com/models/qwen3-next-80b-a3b-instruct-jns) [Prabanta Dev](https://savrn.com/model-publishers/prabanta-dev) On a laptop running Debian 13 Linux, with an Intel Core Ultra 9 185H, its integrated Arc GPU and 32 GB of memory, Janas-LLM reads a prompt at 62 tok/s and writes at 24 tok/s, 37 tok/s with its own draft block in a chat: llama.cpp, on the same laptop, 16 and 2.8. Measured on 9-11 October 2026: llama-bench -p 1024 -n 16 and janas-bench --prompt 1024 --gen 16, both at their best configuration; the drafts with janas-bench on a chat reply, the mean of three rounds (a draft guesses the next tokens and the model checks them: the same text, faster). On other machines the figures will differ. Details: Janas README. - Debian 13 (trixie) Linux, x86-64, kernel 6.12.107 - its integrated Intel Arc GPU… Open weights apache-2.0 [View model](https://savrn.com/models/qwen3-next-80b-a3b-instruct-jns) Model · Text generation ### [qwen3-vl-4b-vietnamese-sft_v2](https://savrn.com/models/qwen3-vl-4b-vietnamese-sft-v2) [Luong Van Phuc](https://savrn.com/model-publishers/phucdu) Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Use the code below to get started with the model. Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - PEFT 0.20.0 Open weights peft [View model](https://savrn.com/models/qwen3-vl-4b-vietnamese-sft-v2) Model · Text generation ### [qwen35_gui_lora](https://savrn.com/models/qwen35-gui-lora) [Jaden Dudik](https://savrn.com/model-publishers/itsnotacreativeuser) Open weights peft [View model](https://savrn.com/models/qwen35-gui-lora) Model · Text generation ### [qwen35a3b-moe-expert-paging-q4](https://savrn.com/models/qwen35a3b-moe-expert-paging-q4) [Max Sun](https://savrn.com/model-publishers/hellosun) EPQ（Expert Paging Quantization）是以 llama.cpp 為基礎的 CPU 推論路徑，讓 22.1 GB 的 Qwen3.6-35B-A3B-UD-Q4KM.gguf 在路由專家權重無法全部放入 RAM 時仍可執行。 GGUF 權重保留在儲存裝置，只有當前選中的專家會分頁載入容量受限的 RAM arena， MoE 專家計算則由 CPU 執行。本 repo 收錄程式碼與量測結果，不包含模型權重。 中文描述名：Qwen3-35B-A3B 專家分頁 4-bit 量化 CPU 推理版。 EPQ 的名稱刻意描述實際機制：MoE expert paging、一般 GGUF 權重量化，以及 CPU 低記憶體推理。本專案不是論文中的 SDQ（Sparse Decomposed Quantization）： 沒有 2:4 結構化稀疏硬體跳零計算，也不宣稱使用 GPU 稀疏加速。sdq/ Python package、sdq C++ 檔案與 SDQ 除錯環境變數是歷史 implementation identifiers， 仍保留以避免破壞既有驗證工具；對外產品名稱與執行入口已改為 EPQ。 需求：Linux x86-64、C++ 工具鏈、CMake、Git、Python 3、足以存放 22.1 GB GGUF 與建置檔的磁碟空間，以及可提供實用解碼速度的本機高速 SSD/NVMe。已測試路徑使用 llama.cpp commit 836d57176dc699a726c55418e4f96b8ca628e1bf。 在其他位置執行時，此腳本會取得本 repo… Open weights apache-2.0 gguf [View model](https://savrn.com/models/qwen35a3b-moe-expert-paging-q4) Model · Text generation ### [qwen38-40b-prune](https://savrn.com/models/qwen38-40b-prune) [DavidB](https://savrn.com/model-publishers/davd-b01) One upstream MoE, six generations of masks, one bit-exact surgery, a rebuilt factual memory, and a three-stage post-training (SFT → SimPO → RLVR) — with every error I made and every idea I falsified along the way. This repository is the complete technical record of the project: the activation-instrumentation artifacts, the mask-generation methodology, the surgery code, the PLE (n-gram memory) rebuild, the trainers, and the full lab notebook. The champion weights live in sibling repos What I set out to do: take Qwen3.8-Flash-Next — an ~180B-class MoE (~130B of core network + 51B of n-gram memory table + 4B MTP) — and cut it by 75% of its experts (512 → 128) while conserving the maximum… Open weights apache-2.0 transformers [View model](https://savrn.com/models/qwen38-40b-prune) Model · Text generation ### [qwen3p5-0p8b-medical-mlp5](https://savrn.com/models/qwen3p5-0p8b-medical-mlp5) [Aditya G](https://savrn.com/model-publishers/adityag-india) Qwen/Qwen3.5-0.8B fine-tuned directly (no LoRA/adapters) on medical reasoning data, to turn a general-purpose model into a domain-focused one. Overfitting to the training data was an accepted trade-off of this experiment; the goal was to study the training/validation loss behaviour. Loss is computed on the assistant response only (reasoning + answer); the prompt is masked. Best validation loss at epoch. See losscurve.png for train vs validation loss. - Research into domain adaptation, partial fine-tuning, and overfitting behaviour of small models. - Offline experimentation with medical-style question answering and reasoning formats. - Clinical decision support, diagnosis, treatment… Open weights apache-2.0 752M parameters 262,144 tokens transformers [View model](https://savrn.com/models/qwen3p5-0p8b-medical-mlp5) Model · Text generation ### [qwen3p6-35b-a3b-2p02bit-r56](https://savrn.com/models/qwen3p6-35b-a3b-2p02bit-r56) [MinJae Choi](https://savrn.com/model-publishers/minjaechoi) Internal research checkpoint. Routed experts average 2.0217 bits; every other weight is BF16. Weights are stored dequantized in BF16 tensors and load with stock transformers / vLLM. License follows the base model. Open weights 35.1B parameters 262,144 tokens transformers [View model](https://savrn.com/models/qwen3p6-35b-a3b-2p02bit-r56) Model · Text generation ### [R1-test4](https://savrn.com/models/r1-test4) [Master Super](https://savrn.com/model-publishers/master103525) Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Use the code below to get started with the model. Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - PEFT 0.18.1 Open weights peft [View model](https://savrn.com/models/r1-test4) Model · Text generation ### [Raon-1-mini-preview](https://savrn.com/models/raon-1-mini-preview) [KRAFTON](https://savrn.com/model-publishers/krafton) Access requested at publisher apache-2.0 28.8B parameters transformers [View model](https://savrn.com/models/raon-1-mini-preview) Model · Text generation ### [RARE-Negative](https://savrn.com/models/rare-negative) [Maggie](https://savrn.com/model-publishers/eurekatian) Open weights apache-2.0 transformers [View model](https://savrn.com/models/rare-negative) Model · Text generation ### [RavenX-CyberAgent-Qwen3.6-35B-A3B-Splash-Q4](https://savrn.com/models/ravenx-cyberagent-qwen3-6-35b-a3b-splash-q4) [Yanni Pang](https://savrn.com/model-publishers/yannipang) RavenX-CyberAgent-Qwen3.6-35B-A3B converted to the Splash schema-4 splash-packed-q4-moe format for Apple Silicon. This repository is a Splash runtime package, not an MLX checkpoint, GGUF file, or Transformers checkpoint. The target and tokenizer come from the which derives from the The compatible DFlash 2 draft and vision weights are reused from All 40 converted language-model layer files were built from RavenX's weights. The reused official Splash weight assets are the draft model and vision encoder. Install Splash on an Apple Silicon Mac, then replace YOURUSERNAME with this repository's owner: After Splash prints Ready, open http://127.0.0.1:8000 or send a chat completion request. The 8K… Open weights apache-2.0 [View model](https://savrn.com/models/ravenx-cyberagent-qwen3-6-35b-a3b-splash-q4) Model · Text generation ### [reachy-mini-motion-planner-0.8b](https://savrn.com/models/reachy-mini-motion-planner-0-8b) [Binh Pham](https://savrn.com/model-publishers/binhpham) Text-to-motion planner for Reachy Mini: a prompt goes in, and a ready-to-play head, antenna and body trajectory comes out. This repo is a complete serving bundle: The response holds the recipe, a one-line idea, and moves: Reachy Mini recorded-move dicts ({"time", "settargetdata": [{"head": 4x4, "antennas", "bodyyaw"}]}), already projected onto the robot's reachable set. Serving effort: "low" runs this model. Speed: ~0.18 s median per prompt on one RTX PRO 6000 (FP8 + fine-tuned MTP; planner 0.14 s), the fastest tier; weaker on multi-phase motions (yawns, bows, build-up events). GPU memory: ~6 GB. Compact system prompt (units, recipe grammar, 4 motion rules, 3 examples), user message = the… Open weights apache-2.0 transformers [View model](https://savrn.com/models/reachy-mini-motion-planner-0-8b) Model · Text generation ### [reachy-mini-motion-planner-27b](https://savrn.com/models/reachy-mini-motion-planner-27b) [Binh Pham](https://savrn.com/model-publishers/binhpham) Text-to-motion planner for Reachy Mini: a prompt goes in, and a ready-to-play head, antenna and body trajectory comes out. This repo is a complete serving bundle: The response holds the recipe, a one-line idea, and moves: Reachy Mini recorded-move dicts ({"time", "settargetdata": [{"head": 4x4, "antennas", "bodyyaw"}]}), already projected onto the robot's reachable set. Serving effort: "high" runs this model. Speed: ~0.84 s median per prompt on one RTX PRO 6000 (FP8 + fine-tuned MTP speculative decoding; planner 0.75 s). GPU memory: ~40 GB with FP8. Compact system prompt (units, recipe grammar, 4 motion rules, 3 examples), user message = the prompt (word. one sentence of context. works… Open weights apache-2.0 transformers [View model](https://savrn.com/models/reachy-mini-motion-planner-27b) Model · Text generation ### [reachy-mini-motion-planner-4b](https://savrn.com/models/reachy-mini-motion-planner-4b) [Binh Pham](https://savrn.com/model-publishers/binhpham) Text-to-motion planner for Reachy Mini: a prompt goes in, and a ready-to-play head, antenna and body trajectory comes out. This repo is a complete serving bundle: The response holds the recipe, a one-line idea, and moves: Reachy Mini recorded-move dicts ({"time", "settargetdata": [{"head": 4x4, "antennas", "bodyyaw"}]}), already projected onto the robot's reachable set. Serving effort: "medium" runs this model. Speed: ~0.31 s median per prompt on one RTX PRO 6000 (FP8 + fine-tuned MTP; planner 0.25 s), 2.7x faster than the 27B. GPU memory: ~12 GB. Compact system prompt (units, recipe grammar, 4 motion rules, 3 examples), user message = the prompt (word. one sentence of context. works best).… Open weights apache-2.0 transformers [View model](https://savrn.com/models/reachy-mini-motion-planner-4b) Model · Text generation ### [reggie-v3.6-rlvr-production-GGUF](https://savrn.com/models/reggie-v3-6-rlvr-production-gguf) [JAMES A BROUSE](https://savrn.com/model-publishers/bigchilli1978) The Reggie open-weights model family, engineered by James Brouse (Bigchilli1978) in Chilliwack, British Columbia, has achieved over 8,521 downloads across Hugging Face: - 7,402 downloads across viral community importance-matrix quantizations by mradermacher (including reggie-v2-7b-i1-GGUF at 4,500+ and reggie-v3-7b-i1-GGUF at 1,760+). - 1,119 direct downloads across 15 official Bigchilli1978 repositories. Notice to Community Quantizers: If you quantize this model (GGUF, EXL2, AWQ, HQQ), feel free to open a discussion or tag Bigchilli1978 so we can link your quants here! Reggie v3.6 is continuously aligned using an on-premises reinforcement learning pipeline (SelfEvolutionDaemon). Rather… Open weights apache-2.0 [View model](https://savrn.com/models/reggie-v3-6-rlvr-production-gguf) Model · Text generation ### [Rem_Mistral_22B](https://savrn.com/models/rem-mistral-22b) [Timur](https://savrn.com/model-publishers/tomry0) This mistral model was trained 2x faster with Unsloth and Huggingface's TRL library. Open weights apache-2.0 22.2B parameters 131,072 tokens transformers [View model](https://savrn.com/models/rem-mistral-22b) Model · Text generation ### [ReScraper](https://savrn.com/models/rescraper) [Chenyan Xiong Research Group at CMU](https://savrn.com/model-publishers/cx-cmu) ReScraper is a 0.6B refiner that replaces the heuristic HTML-to-text stack of a pretraining data pipeline — a rule-based scraper followed by rule-based cleaning filters — with a single small language model. The model reads the visible text of a raw web page, rendered one block per line with line ids, and generates a short program: 1. An block of line removals (rm 1-4, rm 6, …) that strips headers, menus, boilerplate and other non-content lines. 2. One operation tag for the page: An executor applies the program to the page, so extraction and cleaning happen in one model instead of a cascade of separate stages. The system prompt and the executor that applies the generated program are in the… Open weights apache-2.0 [View model](https://savrn.com/models/rescraper) Model · Text generation ### [revonance-drafter-qwen2.5-0.5b](https://savrn.com/models/revonance-drafter-qwen2-5-0-5b) [Yuan Hao Sum](https://savrn.com/model-publishers/trychoosing) The Drafter agent of a LangGraph multi-agent support system for the fictional shop ReVonance. Base model Qwen/Qwen2.5-0.5B-Instruct, fine-tuned with LoRA (merged). Stage: sft (sft = supervised on the response simulator; dpo = then aligned with human and critic preferences via Direct Preference Optimisation). Knowledge-base fingerprint: f9cdb07b5df6. Demo model trained on synthetic data; it only knows the toy policies of ReVonance. Open weights apache-2.0 494M parameters 32,768 tokens transformers [View model](https://savrn.com/models/revonance-drafter-qwen2-5-0-5b) Model · Text generation ### [revonance-drafter-smollm2-360m](https://savrn.com/models/revonance-drafter-smollm2-360m) [Yuan Hao Sum](https://savrn.com/model-publishers/trychoosing) The Drafter agent of a LangGraph multi-agent support system for the fictional shop ReVonance. Base model HuggingFaceTB/SmolLM2-360M-Instruct, fine-tuned with LoRA (merged). Stage: sft (sft = supervised on the response simulator; dpo = then aligned with human and critic preferences via Direct Preference Optimisation). Knowledge-base fingerprint: f9cdb07b5df6. Demo model trained on synthetic data; it only knows the toy policies of ReVonance. Open weights apache-2.0 362M parameters 8,192 tokens transformers [View model](https://savrn.com/models/revonance-drafter-smollm2-360m) Model · Text generation ### [rick-morty-distilgpt2](https://savrn.com/models/rick-morty-distilgpt2) [Zune toka](https://savrn.com/model-publishers/hfzune) This model is a fine-tuned version of distilbert/distilgpt2 on the None dataset. It achieves the following results on the evaluation set: The following hyperparameters were used during training: - learningrate: 5e-05 - trainbatchsize: 4 - evalbatchsize: 4 - gradientaccumulationsteps: 4 - totaltrainbatchsize: 16 - lrschedulertype: linear - numepochs: 1 - mixedprecisiontraining: Native AMP - Transformers 5.17.0 - Pytorch 2.11.0+cu128 - Datasets 5.0.1 - Tokenizers 0.23.1 Open weights apache-2.0 82M parameters transformers [View model](https://savrn.com/models/rick-morty-distilgpt2) Model · Text generation ### [RK182X-LLM-glm-edge-1.5b-chat](https://savrn.com/models/rk182x-llm-glm-edge-1-5b-chat) [RKNNAI](https://savrn.com/model-publishers/rknnai) 本仓库提供由 THUDM/glm-edge-1.5b-chat 转换得到的 RKNN LLM 模型。 - Model ID：RKNNAI/RK182X-LLM-glm-edge-1.5b-chat - 模型显示名称：RK182X-LLM-glm-edge-1.5b-chat - 源模型：THUDM/glm-edge-1.5b-chat - 模型类型：LLM ModelScope 完整下载： Hugging Face 完整下载： ModelScope 指定配置下载： Hugging Face 指定配置下载： - 使用配套 RKNN Runtime 和驱动；运行前用 rknn-smi -v 检查设备端版本。 - 需使用配套 chat template 和推理 Demo。 源模型许可证见根目录 LICENSE；同时遵守 RKNN Toolkit 和 RKNN Runtime 许可条款。 本发布物为源模型的 RKNN 转换/量化衍生物，保留原始许可与归属声明，见 Notice。 源模型使用 GLM-Edge License，并非 Apache 2.0。 Built with GLM-Edge. Open weights other [View model](https://savrn.com/models/rk182x-llm-glm-edge-1-5b-chat) Model · Text generation ### [RK182X-LLM-HY-MT1.5-1.8B](https://savrn.com/models/rk182x-llm-hy-mt1-5-1-8b) [RKNNAI](https://savrn.com/model-publishers/rknnai) 本仓库提供由 tencent/HY-MT1.5-1.8B 转换得到的 RKNN LLM 模型。 - Model ID：RKNNAI/RK182X-LLM-HY-MT1.5-1.8B - 模型显示名称：RK182X-LLM-HY-MT1.5-1.8B - 源模型：tencent/HY-MT1.5-1.8B - 模型类型：LLM ModelScope 完整下载： Hugging Face 完整下载： ModelScope 指定配置下载： Hugging Face 指定配置下载： - 使用配套 RKNN Runtime 和驱动；运行前用 rknn-smi -v 检查设备端版本。 源模型许可证见根目录 LICENSE；同时遵守 RKNN Toolkit 和 RKNN Runtime 许可条款。 本发布物为源模型的 RKNN 转换/量化衍生物，保留原始许可与归属声明，见 Notice。 源模型使用 TENCENT HY COMMUNITY LICENSE AGREEMENT，并非 Apache 2.0。使用与再分发受 LICENSE 第 5 节及附录限制；腾讯与本转换发布方不存在隶属、赞助或背书关系。对外提供服务时须依条款披露实际服务主体。 Open weights other [View model](https://savrn.com/models/rk182x-llm-hy-mt1-5-1-8b) Model · Text generation ### [RK182X-LLM-MiniCPM5-1B](https://savrn.com/models/rk182x-llm-minicpm5-1b) [RKNNAI](https://savrn.com/model-publishers/rknnai) 本仓库提供由 openbmb/MiniCPM5-1B 转换得到的 RKNN LLM 模型。 - Model ID：RKNNAI/RK182X-LLM-MiniCPM5-1B - 模型显示名称：RK182X-LLM-MiniCPM5-1B - 源模型：openbmb/MiniCPM5-1B - 模型类型：LLM ModelScope 完整下载： Hugging Face 完整下载： ModelScope 指定配置下载： Hugging Face 指定配置下载： - 使用配套 RKNN Runtime 和驱动；运行前用 rknn-smi -v 检查设备端版本。 源模型许可证见根目录 LICENSE；同时遵守 RKNN Toolkit 和 RKNN Runtime 许可条款。 本发布物为源模型的 RKNN 转换/量化衍生物，保留原始许可与归属声明，见 Notice。 Open weights apache-2.0 [View model](https://savrn.com/models/rk182x-llm-minicpm5-1b) Model · Text generation ### [RK182X-LLM-MiniCPM5-2B](https://savrn.com/models/rk182x-llm-minicpm5-2b) [RKNNAI](https://savrn.com/model-publishers/rknnai) 本仓库提供由 openbmb/MiniCPM5-2B 转换得到的 RKNN LLM 模型。 - Model ID：RKNNAI/RK182X-LLM-MiniCPM5-2B - 模型显示名称：RK182X-LLM-MiniCPM5-2B - 源模型：openbmb/MiniCPM5-2B - 模型类型：LLM ModelScope 完整下载： Hugging Face 完整下载： ModelScope 指定配置下载： Hugging Face 指定配置下载： - 使用配套 RKNN Runtime 和驱动；运行前用 rknn-smi -v 检查设备端版本。 源模型许可证见根目录 LICENSE；同时遵守 RKNN Toolkit 和 RKNN Runtime 许可条款。 本发布物为源模型的 RKNN 转换/量化衍生物，保留原始许可与归属声明，见 Notice。 Open weights apache-2.0 [View model](https://savrn.com/models/rk182x-llm-minicpm5-2b) Model · Text generation ### [RK182X-LLM-Qwen2.5-0.5B-Instruct](https://savrn.com/models/rk182x-llm-qwen2-5-0-5b-instruct) [RKNNAI](https://savrn.com/model-publishers/rknnai) 本仓库提供由 Qwen/Qwen2.5-0.5B-Instruct 转换得到的 RKNN LLM 模型。 - Model ID：RKNNAI/RK182X-LLM-Qwen2.5-0.5B-Instruct - 模型显示名称：RK182X-LLM-Qwen2.5-0.5B-Instruct - 源模型：Qwen/Qwen2.5-0.5B-Instruct - 模型类型：LLM ModelScope 完整下载： Hugging Face 完整下载： ModelScope 指定配置下载： Hugging Face 指定配置下载： - 使用配套 RKNN Runtime 和驱动；运行前用 rknn-smi -v 检查设备端版本。 源模型许可证见根目录 LICENSE；同时遵守 RKNN Toolkit 和 RKNN Runtime 许可条款。 本发布物为源模型的 RKNN 转换/量化衍生物，保留原始许可与归属声明，见 Notice。 Open weights apache-2.0 [View model](https://savrn.com/models/rk182x-llm-qwen2-5-0-5b-instruct) ## Questions ### Which Text generation models are most downloaded? By monthly downloads reported by the Hugging Face Hub: . ## Other Tasks - [Image and Text to Text](https://savrn.com/models/tasks/image-and-text-to-text) 577 - [Feature Extraction](https://savrn.com/models/tasks/feature-extraction) 398 - [Robotics](https://savrn.com/models/tasks/robotics) 398 - [Text Classification](https://savrn.com/models/tasks/text-classification) 331 - [Text to Image](https://savrn.com/models/tasks/text-to-image) 232 - [Speech Recognition](https://savrn.com/models/tasks/speech-recognition) 187 - [Text to Speech](https://savrn.com/models/tasks/text-to-speech) 145 - [Image Classification](https://savrn.com/models/tasks/image-classification) 139 - [Audio Classification](https://savrn.com/models/tasks/audio-classification) 130 - [Sentence Similarity](https://savrn.com/models/tasks/sentence-similarity) 103 - [Reinforcement Learning](https://savrn.com/models/tasks/reinforcement-learning) 100 - [Token Classification](https://savrn.com/models/tasks/token-classification) 100 [See all](https://savrn.com/models/tasks)