THE MODEL DIRECTORY

Explore models.

MY HARDWARE

Your machine

Choose your actual machine to personalize model and workflow ratings.

Choose my machine

GPU maker

Find your GPUSearch the full catalog or browse by GPU generation below.GPU modelRTX 3060 desktop is available in 12 GB and 8 GB versions; pick the memory printed on your card.System RAM (GB) Optional; helps assess CPU offloadOperating systemCPU (optional)

Saved only in this browser. Model cards show a machine-specific rating when the setup supports one, or explain what evidence is still needed.

TASK → MACHINE → MODEL

Find your fit

What do you want to do?Model variantContextExperience

Choose a specific task to see its recommended model and documented alternatives. A memory-fit estimate is distinct from a verified run setup. Estimated ratings name their package and assumptions; source-only cards link to the available run path without inventing a hardware verdict.

Model freshnessModel popularityArena categoryArena · Open Source rankHow scores work ↗

Freshness uses release date, or repository age when no release date is available. Rank filters use the selected Arena category, with #1 ranked highest. Unranked models remain visible unless you filter them out. Hardware and score filters must match the same version.

More filters

PublisherLicenseCapabilityDeployment

37 modelsSorted by release recency, documented run path, and source quality.

MiMo V2.6

Explore 3 versions of MiMo V2.6. Compare model sizes, packages, hardware requirements, and available setups.

3 versionsCompare sizes & formats

ShowingMiMo V2.6 Distill Qwen 9B

Freshness 83 out of 100. Repository age: 48 days old, Sep 21, 2026. Measures age, not quality.

Popularity 60 out of 100. 17,769 downloads in 30 days; 694 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE9.41B

LOCALRENTED GPU

Compare versions →

Xing4.0

Explore 3 versions of Xing4.0. Compare model sizes, packages, hardware requirements, and available setups.

3 versionsCompare sizes & formats

ShowingXing4.0 29B A4B

Freshness 82 out of 100. Repository age: 53 days old, Sep 16, 2026. Measures age, not quality.

Popularity 56 out of 100. 5,343 downloads in 30 days; 1,865 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE31.22B

LOCALRENTED GPU

Compare versions →

Swift Qwen3.8 27b

Swift Qwen3.8 27b is a fine-tuned derivative of Qwen/Qwen3.8-27B, released by ukisai under the swift-open-license-1.0. It is an image-text-to-text model with safetensors weights and reasoning tags. According to its official model card, it is described as a reasoning-efficient variant that uses 58.3% fewer thinking tokens while maintaining near-identical performance (<1% loss) and achieving a x1.95 speed-up on several tasks. The training approach involved identifying reasoning-marker tokens that appear to trigger overthinking in Qwen's reasoning rollouts, then fine-tuning by penalizing those tokens while reasoning. It also includes a transfer component derived from BottleCap AI's ThinkingCap-Qwen3.6-27B. Evaluation results are stated to compare the Qwen3.8-27B BF16 base with the same base plus the Swift adapter.

Freshness 79 out of 100. Repository age: 61 days old, Sep 8, 2026. Measures age, not quality.

Popularity 50 out of 100. 2,619 downloads in 30 days; 639 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE27.78B

LOCALRENTED GPU

Repository opened Sep 8, 2026

MiniCPM5

Explore 10 versions of MiniCPM5. Compare model sizes, packages, hardware requirements, and available setups.

10 versionsCompare sizes & formats

ShowingMiniCPM5 2B GPTQ

Freshness 79 out of 100. Repository age: 62 days old, Sep 7, 2026. Measures age, not quality.

Popularity 40 out of 100. 965 downloads in 30 days; 22 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE2.52B

LOCALRENTED GPU

Compare versions →

NeoHorse 1

Explore 2 versions of NeoHorse 1. Compare model sizes, packages, hardware requirements, and available setups.

2 versionsCompare sizes & formats

ShowingNeoHorse 1 4B

Freshness 78 out of 100. Repository age: 64 days old, Sep 5, 2026. Measures age, not quality.

Popularity 54 out of 100. 3,396 downloads in 30 days; 2,556 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE4.21B

LOCALRENTED GPU

Compare versions →

K2 Horizon

Explore 2 versions of K2 Horizon. Compare model sizes, packages, hardware requirements, and available setups.

2 versionsCompare sizes & formats

ShowingK2 Horizon 7B

Freshness 77 out of 100. Repository age: 68 days old, Sep 1, 2026. Measures age, not quality.

Popularity 66 out of 100. 84,383 downloads in 30 days; 282 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE9.00B

LOCALRENTED GPU

Compare versions →

GLM 5.3 CYBERSECURITY FP8

GLM 5.3 CYBERSECURITY FP8 is a text-generation model published at dealignai/GLM-5.3-CYBERSECURITY-FP8 under the MIT license. It is described as a cybersecurity-domain, refusal-reduced ('abliterated'/'crack') FP8 variant based on zai-org/GLM-5.3 and JANGQ-AI/GLM-5.3-FP8. The model card states it targets offensive-security, red-team, exploit-dev, reverse-engineering, evasion, phishing, credential-attack, and malware-analysis content, with refusal reduction noted as domain-specific rather than general-purpose. It is a 753B total parameter, text-only glm_moe_dsa MoE model with 78 layers, using routed FP8 experts with edited bf16 residual writers, and is intended for native FP8 tensor-core speed on Hopper GPUs. Supported languages listed are en, zh, ru, sr, hi, fr, es, ar, ko, and ja.

Freshness 76 out of 100. Repository age: 70 days old, Aug 30, 2026. Measures age, not quality.

Popularity 70 out of 100. 127,638 downloads in 30 days; 881 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE753.33B

LOCALRENTED GPU

Repository opened Aug 30, 2026

Bonsai 2 · 27B

A compressed 27B-class reasoning model with publisher-provided GGUF packs and a dedicated llama.cpp fork for CUDA, Metal, and CPU.

Freshness 82 out of 100. Release age: 52 days old, Sep 17, 2026. Measures age, not quality.

Popularity 85 out of 100. 1,710,588 downloads in 30 days; 2,952 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE27.36B

LOCALRENTED GPU

Released Sep 17, 2026

GLM 5.3

Explore 2 versions of GLM 5.3. Compare model sizes, packages, hardware requirements, and available setups.

2 versionsCompare sizes & formats

ShowingGLM-5.3 Flash

Freshness 75 out of 100. Repository age: 75 days old, Aug 25, 2026. Measures age, not quality.

Popularity 92 out of 100. 6,423,926 downloads in 30 days; 3,083 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

Arena Open Source Agents: rank 10 of 23. Ranked model: GLM-5.3-Flash. Export dated Nov 4, 2026. Source checked Nov 7, 2026. Human preference ranking, not hardware suitability.+3

MODEL SIZE321B total · 18B active

LOCALRENTED GPU

Compare versions →

Spark X2.5 4B

Spark X2.5 4B is a text-generation model from the repository XHToken/Spark-X2.5-4B, licensed under apache-2.0 and compatible with the Hugging Face Transformers format. The model card describes it as a compact, general-purpose language model that uses a hybrid attention architecture (one full-attention layer and three sliding-window attention layers) and natively supports context windows up to 1M tokens. It is presented as supporting conversation, writing, translation, reasoning, coding, tool use, and agentic workflows, with broad language coverage and compatibility with several inference frameworks and platforms. The model was trained on Huawei Ascend clusters using large-scale reinforcement learning and post-training techniques. No specific performance numbers or hardware requirements are stated in the provided source.

Freshness 75 out of 100. Repository age: 76 days old, Aug 24, 2026. Measures age, not quality.

Popularity 64 out of 100. 29,150 downloads in 30 days; 1,440 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE4.11B

LOCALRENTED GPU

Repository opened Aug 24, 2026

Ornith 1.5

Explore 3 versions of Ornith 1.5. Compare model sizes, packages, hardware requirements, and available setups.

3 versionsCompare sizes & formats

ShowingOrnith 1.5 35B A3B

Freshness 73 out of 100. Repository age: 82 days old, Aug 18, 2026. Measures age, not quality.

Popularity 74 out of 100. 285,458 downloads in 30 days; 781 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE35.95B

LOCALRENTED GPU

Compare versions →

DeepSeek V4

Explore 5 versions of DeepSeek V4. Compare model sizes, packages, hardware requirements, and available setups.

5 versionsCompare sizes & formats

ShowingDeepSeek V4 Pro 0813

Freshness 72 out of 100. Repository age: 87 days old, Aug 13, 2026. Measures age, not quality.

Popularity 71 out of 100. 145,382 downloads in 30 days; 912 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE1650.50B

LOCALRENTED GPU

Compare versions →

Kolibri 1 · BF16

An open-weight English/German reasoning and tool-calling model with a publisher-maintained vLLM plugin and multi-GPU serving instructions.

Freshness 87 out of 100. Release age: 36 days old, Oct 3, 2026. Measures age, not quality.

Popularity 48 out of 100. 3,340 downloads in 30 days; 69 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE78B total · 3.46B active

LOCALRENTED GPU

Released Oct 3, 2026

LFM2.5

Explore 3 versions of LFM2.5. Compare model sizes, packages, hardware requirements, and available setups.

3 versionsCompare sizes & formats

ShowingLFM2.5 2.6B

Freshness 67 out of 100. Repository age: 103 days old, Jul 28, 2026. Measures age, not quality.

Popularity 66 out of 100. 53,680 downloads in 30 days; 860 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE2.70B

LOCALRENTED GPU

Compare versions →

Signal 3.8 27B GGUF

Signal 3.8 27B GGUF is a minimally invasive fine-tune of Qwen3.8-27B, published as a GGUF-format model for the image-text-to-text pipeline under the apache-2.0 license. It is designed for lower generation latency and better token efficiency by producing more direct answers, with fewer preambles, excessive formatting, sign-offs, and explanatory narration. The model is trained by self-distillation on Qwen3.8-27B's own answers and is presented as a drop-in GGUF replacement for llama.cpp setups that already run Qwen3.8-27B.

Freshness 79 out of 100. Repository age: 60 days old, Sep 9, 2026. Measures age, not quality.

Popularity 56 out of 100. 15,082 downloads in 30 days; 105 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

LOCALRENTED GPU

Repository opened Sep 9, 2026

Hy3

Hy3 from tencent. Source-based hardware guidance from its published configuration.

Freshness 61 out of 100. Repository age: 129 days old, Jul 2, 2026. Measures age, not quality.

Popularity 81 out of 100. 1,182,445 downloads in 30 days; 990 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

Arena Open Source Text: rank 14 of 228. Ranked model: Hy3. Export dated Nov 6, 2026. Source checked Nov 7, 2026. Human preference ranking, not hardware suitability.+2

MODEL SIZE298.79B

LOCALRENTED GPU

Repository opened Jul 2, 2026

Qwen3.8

Explore 4 versions of Qwen3.8. Compare model sizes, packages, hardware requirements, and available setups.

4 versionsCompare sizes & formats

ShowingQwen3.8 27B GGUF

Freshness 73 out of 100. Repository age: 82 days old, Aug 18, 2026. Measures age, not quality.

Popularity 81 out of 100. 2,094,837 downloads in 30 days; 205 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

Arena Open Source Text: rank 26 of 228. Ranked model: Qwen3.8-27B. Original model ranking, not a benchmark of this package. Export dated Nov 6, 2026. Source checked Nov 7, 2026. Human preference ranking, not hardware suitability.+3

LOCALRENTED GPU

Compare versions →

GLM 5.2

Explore 2 versions of GLM 5.2. Compare model sizes, packages, hardware requirements, and available setups.

2 versionsCompare sizes & formats

Freshness 57 out of 100. Repository age: 145 days old, Jun 16, 2026. Measures age, not quality.

Popularity 78 out of 100. 318,480 downloads in 30 days; 5,204 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE753.33B

LOCALRENTED GPU

Compare versions →

Step 3.7 Flash

Step 3.7 Flash is a sparse Mixture-of-Experts (MoE) vision-language model with 198B total parameters, combining a 196B-parameter language backbone and a 1.8B-parameter vision encoder. It is licensed under Apache 2.0, uses the image-text-to-text pipeline, and is associated with the safetensors format and Transformers library. The model activates approximately 11B parameters per token and is described as supporting a 256k context window with three selectable reasoning levels (low, medium, high). It is intended for high-frequency production workloads and agentic workflows involving perception, search, and reasoning.

Freshness 52 out of 100. Repository age: 169 days old, May 23, 2026. Measures age, not quality.

Popularity 57 out of 100. 10,428 downloads in 30 days; 463 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE201.37B

LOCALRENTED GPU

Repository opened May 23, 2026

Bonsai

Explore 2 versions of Bonsai. Compare model sizes, packages, hardware requirements, and available setups.

2 versionsCompare sizes & formats

ShowingBonsai 27B · original 1-bit

Freshness 61 out of 100. Repository age: 127 days old, Jul 4, 2026. Measures age, not quality.

Popularity 73 out of 100. 222,350 downloads in 30 days; 941 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

LOCALRENTED GPU

Compare versions →

Granite Embedding Multilingual R2

Explore 2 versions of Granite Embedding Multilingual R2. Compare model sizes, packages, hardware requirements, and available setups.

2 versionsCompare sizes & formats

Showinggranite embedding 311m multilingual r2

Freshness 46 out of 100. Repository age: 202 days old, Apr 20, 2026. Measures age, not quality.

Popularity 65 out of 100. 88,829 downloads in 30 days; 145 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE0.31B

LOCALRENTED GPU

Compare versions →

Nemotron 3 Nano Omni 30B A3B Reasoning BF16

NVIDIA Nemotron 3 Nano Omni 30B A3B Reasoning BF16 is a multimodal large language model described in its official model card as a 31B-parameter (approximately 3B active per token) Mamba2-Transformer hybrid Mixture-of-Experts model. It accepts video, audio, image, and text inputs and produces text outputs. It supports a maximum context of 256k tokens, has a reasoning mode enabled by default that can be toggled via enable_thinking, and is released under the NVIDIA Open Model Agreement as an any-to-any pipeline model using safetensors. The card lists BF16, FP8, and NVFP4 formats and various GPU suggestions, including minimum BF16 single-GPU use with 1× H100 80GB. The model is positioned for enterprise multimodal understanding workflows such as Q&A, summarization, transcription, document intelligence, and GUI/agentic automation. No testing, performance benchmarks, or compatibility claims beyond the provided card are inferred here.

Freshness 46 out of 100. Repository age: 202 days old, Apr 20, 2026. Measures age, not quality.

Popularity 69 out of 100. 130,604 downloads in 30 days; 442 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE33.02B

LOCALRENTED GPU

Repository opened Apr 20, 2026

Mellum2.1 Thinking

A 12B-total, 2.5B-active open-weight thinking model for coding and tool use. Start with JetBrains’ official GGUF on a suitable desktop, or serve the BF16 checkpoint with vLLM on a larger GPU.

Freshness 83 out of 100. Release age: 49 days old, Sep 20, 2026. Measures age, not quality.

Popularity 55 out of 100. 11,306 downloads in 30 days; 193 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE12B total · 2.5B active

LOCALRENTED GPU

Released Sep 20, 2026

DeepSeek V4.1 Flash

An open-weight model for reasoning, coding, agents, and image understanding. DeepSeek V4.1 Flash combines a 552B backbone with sparse activation and up to one million tokens of context. Choose the original checkpoint or a community uncensored variant below.

Freshness 80 out of 100. Repository age: 59 days old, Sep 10, 2026. Measures age, not quality.

Popularity 87 out of 100. 1,911,780 downloads in 30 days; 4,887 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE552B backbone · 8B active prefill / 16B active decode

LOCALRENTED GPU

Repository opened Sep 10, 2026

Kimi K3

An open-weight multimodal model for long-running coding, reasoning, and knowledge work.

Freshness 57 out of 100. Repository age: 148 days old, Jun 13, 2026. Measures age, not quality.

Popularity 84 out of 100. 771,618 downloads in 30 days; 11,940 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

Arena Open Source Agents: rank 5 of 23. Ranked model: Kimi K3 (Max). Export dated Nov 4, 2026. Source checked Nov 7, 2026. Human preference ranking, not hardware suitability.+2

MODEL SIZE2.8T total · 104B active

LOCALRENTED GPU

Repository opened Jun 13, 2026

Qwen3.5

Explore 3 versions of Qwen3.5. Compare model sizes, packages, hardware requirements, and available setups.

3 versionsCompare sizes & formats

ShowingQwen3.5 · 35B A3B

Freshness 37 out of 100. Repository age: 257 days old, Feb 24, 2026. Measures age, not quality.

Popularity 84 out of 100. 1,621,768 downloads in 30 days; 1,555 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

Arena Open Source Web coding: rank 58 of 62. Ranked model: Qwen3.5-35B-A3B. Export dated Nov 6, 2026. Source checked Nov 7, 2026. Human preference ranking, not hardware suitability.+1

MODEL SIZE35B total · 3B active

LOCALRENTED GPU

Compare versions →

Google DeepMind

Gemma 4

Explore 4 versions of Gemma 4. Compare model sizes, packages, hardware requirements, and available setups.

4 versionsCompare sizes & formats

ShowingGemma 4 · 12B

Freshness 52 out of 100. Repository age: 169 days old, May 23, 2026. Measures age, not quality.

Popularity 85 out of 100. 1,846,252 downloads in 30 days; 1,725 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

LOCALRENTED GPU

Compare versions →

GLM 5

GLM-5 is a text-generation model released by zai-org under the MIT license. It targets complex systems engineering and long-horizon agentic tasks. The model card states that GLM-5 scales to 744B parameters (40B active), increases pre-training data to 28.5T tokens compared with GLM-4.5, and integrates DeepSeek Sparse Attention (DSA). It supports English and Chinese, uses the transformers library, and is available in safetensors format. The model card also lists benchmark results and references API services, a technical blog, and a community.

Freshness 35 out of 100. Repository age: 270 days old, Feb 11, 2026. Measures age, not quality.

Popularity 84 out of 100. 1,449,282 downloads in 30 days; 2,141 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

Arena Open Source Text: rank 12 of 228. Ranked model: GLM-5. Export dated Nov 6, 2026. Source checked Nov 7, 2026. Human preference ranking, not hardware suitability.+1

MODEL SIZE753.86B

LOCALRENTED GPU

Repository opened Feb 11, 2026

Qwen3-Coder-Next · 80B A3B

Qwen3-Coder-Next is an open-weight text-generation model from Qwen/Qwen3-Coder-Next, licensed under apache-2.0. It is designed specifically for coding agents and local development. The model has 80B total parameters with 3B activated parameters, a 262,144-token native context length, and a hybrid architecture combining Gated DeltaNet, Gated Attention, and Mixture-of-Experts layers. It supports only non-thinking mode and does not generate thinking blocks.

Freshness 34 out of 100. Repository age: 282 days old, Jan 30, 2026. Measures age, not quality.

Popularity 77 out of 100. 428,592 downloads in 30 days; 1,735 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE80B total / 3B active

LOCALRENTED GPU

Repository opened Jan 30, 2026

Mistral Small 4 · 119B

Mistral Small 4 119B A6B is an Apache-2.0 licensed hybrid model from mistralai that combines general instruction, reasoning, and Devstral-style capabilities in one model. It uses a Mixture-of-Experts architecture with 119B total parameters, 6.5B activated per token, 128 experts with 4 active, and a 256k context length. It supports multimodal input (text and image) with text output, multilingual support across dozens of languages, system prompts, function calling, JSON output, and configurable reasoning effort. The model card also mentions optional efficiency resources: an eagle head for speculative decoding and an NVFP4 checkpoint. It is tagged for vLLM and offered in safetensors format.

Freshness 33 out of 100. Repository age: 289 days old, Jan 23, 2026. Measures age, not quality.

Popularity 62 out of 100. 29,619 downloads in 30 days; 472 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE119B total · 6.5B active

LOCALRENTED GPU

Repository opened Jan 23, 2026

GLM 4.7

Explore 2 versions of GLM 4.7. Compare model sizes, packages, hardware requirements, and available setups.

2 versionsCompare sizes & formats

ShowingGLM 4.7 Flash

Freshness 32 out of 100. Repository age: 293 days old, Jan 19, 2026. Measures age, not quality.

Popularity 82 out of 100. 1,016,425 downloads in 30 days; 1,895 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

Arena Open Source Text: rank 78 of 228. Ranked model: GLM-4.7-Flash. Export dated Nov 6, 2026. Source checked Nov 7, 2026. Human preference ranking, not hardware suitability.

MODEL SIZE31.22B

LOCALRENTED GPU

Compare versions →

Stable DiffCoder 8B Instruct

Stable-DiffCoder-8B-Instruct is a code-focused mask diffusion language model by ByteDance-Seed, built on the Seed-Coder architecture and fine-tuned from Stable-DiffCoder-8B-Base. It uses block diffusion continual pretraining followed by supervised fine-tuning and is instruction-tuned for alignment with user intent. The model card lists an 8K context length, text-generation pipeline, transformers library support, MIT license, and safetensors format.

Freshness 32 out of 100. Repository age: 297 days old, Jan 15, 2026. Measures age, not quality.

Popularity 41 out of 100. 722 downloads in 30 days; 143 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE8.25B

LOCALRENTED GPU

Repository opened Jan 15, 2026

dealignai · based on Z.ai GLM-5.3

GLM-5.3 UNCENSORED · FP8

A downloadable, community-modified GLM-5.3 checkpoint with reduced refusal behavior. Run the exact FP8 weights on an eight-H200 Linux server and connect a PC harness through a private API, or try a clearly identified different hosted GLM-5.3 derivative first.

Freshness 76 out of 100. Repository age: 70 days old, Aug 30, 2026. Measures age, not quality.

Popularity 49 out of 100. 3,193 downloads in 30 days; 207 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE753B total, according to the derivative’s model card

LOCALRENTED GPU

Repository opened Aug 30, 2026

Ministral 3 14B Reasoning 2512

Ministral 3 14B Reasoning 2512 is the reasoning post-trained version of the largest model in the Ministral 3 family, released by mistralai under the apache-2.0 license. It combines a 13.5B language model with a 0.4B vision encoder, supports vision, multilingual input across eleven listed languages, system prompts, native function calling and JSON outputting, and a 256k context window. The model is described as edge-optimized and is said to fit in 32GB of VRAM in BF16, or less than 24GB of RAM/VRAM when quantized. It is intended for reasoning-heavy tasks such as math, coding, and STEM, as well as private or local deployments, agentic uses, and fine-tuning. The model card recommends using the provided system prompt, keeping reasoning traces in context for multi-turn interactions, a temperature of 1 for most environments, limiting the number of tools, and maintaining an aspect ratio close to 1:1 for images.

Freshness 24 out of 100. Repository age: 373 days old, Oct 31, 2025. Measures age, not quality.

Popularity 72 out of 100. 329,032 downloads in 30 days; 176 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE13.95B

LOCALRENTED GPU

Repository opened Oct 31, 2025

Mistral Large 3 675B Instruct 2512 NVFP4

Mistral Large 3 675B Instruct 2512 NVFP4 is an Apache-2.0 licensed, instruct post-trained multimodal granular Mixture-of-Experts model. It has 675B total parameters with 41B active parameters, comprising a 673B-parameter MoE language model with 39B active and a 2.5B vision encoder. It is an NVFP4 post-training-activation quantized checkpoint derived from Mistral-Large-3-675B-Instruct-2512 and created with llm-compressor in collaboration with vLLM and Red Hat. The model card reports support for vision, multilingual use, system prompts, agentic capabilities with native function calling and JSON output, and a 256k context window. It is described as deployable on-premises in NVFP4 on a single node of H100s or A100s, while the FP8 version is described as deployable on a single node of B200s or H200s.

Freshness 26 out of 100. Repository age: 345 days old, Nov 28, 2025. Measures age, not quality.

Popularity 51 out of 100. 5,821 downloads in 30 days; 97 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

LOCALRENTED GPU

Repository opened Nov 28, 2025

Qwen2.5 Coder · 7B

Code generation, repair, and reasoning in a compact open-weight coding model.

Freshness 5 out of 100. Repository age: 782 days old, Sep 17, 2024. Measures age, not quality.

Popularity 82 out of 100. 1,302,537 downloads in 30 days; 865 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE7.61B

LOCALRENTED GPU

Verified source

Qwen base · Strata runtime · ISTA-DASLab Coder derivative

Qwen3.8-Flash-Next · Strata

Run Qwen3.8-Flash-Next through Strata, an inference runtime that shares model experts across GPU memory, system RAM, and storage. Start with its reduced-expert Coder derivative for a low-VRAM coding assistant, or choose a fuller model pack when you have more RAM.

Freshness 75 out of 100. Repository age: 76 days old, Aug 24, 2026. Measures age, not quality.

Popularity 90 out of 100. 3,607,555 downloads in 30 days; 6,834 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

MODEL SIZE125B base · ~6B active; Coder removes half the experts

LOCALRENTED GPU

Partially field-tested