# SKILLRET-Reranker-0.6B by ThakiCloud: Open-Weight Model Source: https://savrn.com/models/skillret-reranker-0-6b-2 Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- ## Runs On What it takes to serve SKILLRET-Reranker-0.6B (596M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them. | Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits | | --- | --- | --- | --- | --- | --- | | 16-bit | 1.2 GB | 1.4 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 8-bit | 0.6 GB | 0.7 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 4-bit | 0.3 GB | 0.4 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the [SAVRN Index](https://savrn.com/ai-index/pricing/gpus), read Oct 7, 2026. [SKILLRET-Reranker-0.6B on every accelerator the SAVRN Index prices, at every precision](https://savrn.com/models/skillret-reranker-0-6b-2/gpus) ## Model Card By ThakiCloud, published under apache-2.0, revision ce09c883af78. This is a reranker fine-tuned for AI agent skill retrieval. Given a natural-language user request and a candidate agent skill, it scores how relevant and useful the skill is for the request. It is designed as the second stage after a first-stage retriever such as [SkillRet-Embedding-0.6B](https://huggingface.co/ThakiCloud/SKILLRET-Embedding-0.6B) or [SkillRet-Embedding-8B](https://huggingface.co/ThakiCloud/SKILLRET-Embedding-8B). The model is fine-tuned from [Qwen/Qwen3-Reranker-0.6B](https://savrn.com/models/qwen3-reranker-0-6b) on the SkillRet benchmark training split with binary cross-entropy on the yes/no token probability. It keeps the scoring interface of Qwen3-Reranker. Technical report: [SkillRet: A Large-Scale Benchmark for Skill Retrieval in LLM Agents (arXiv:2605.05726)](https://arxiv.org/abs/2605.05726) ### Usage #### Transformers [Read the full model card (603 words)](https://savrn.com/models/skillret-reranker-0-6b-2/card) ## Configuration Architecture Qwen3ForCausalLM Context length (tokens) 40,960 Layers 28 Hidden size 1,024 Feed-forward size 3,072 Attention heads 16 Key/value heads 8 Head dimension 128 Vocabulary size 151,669 Model type qwen3 ## Identity and Version Repository ThakiCloud/SKILLRET-Reranker-0.6B Publisher ThakiCloud Task Text ranking Modality Other Library transformers Parameters 596M parameters Languages en Revision ce09c883af78affc72a85014e040f3d28a5b1711 First published 2026-09-22 Last updated 2026-09-24 ## Files and Weights 8 files, 1.2 GB in total. The weights are 1 file totalling 1.2 GB in safetensors. Weights1 file · 1.2 GB Configuration2 files · 1.6 KB Tokenizer2 files · 11.4 MB Documentation1 file · 5.9 KB Other1 file · 741 B Repository1 file · 1.6 KB Every file | File | Type | Size | SHA-256 | | --- | --- | --- | --- | | model.safetensors | Weights | 1.2 GB | ace7f2068c57 | | config.json | Configuration | 1.4 KB | — | | generation_config.json | Configuration | 187 B | — | | README.md | Documentation | 5.9 KB | — | | chat_template.jinja | Other | 741 B | — | | .gitattributes | Repository | 1.6 KB | — | | tokenizer.json | Tokenizer | 11.4 MB | be75606093db | | tokenizer_config.json | Tokenizer | 375 B | — | ## License and Download License apache-2.0 Access Open weights, no gate Download size 1.2 GB [Download from ThakiCloud](https://huggingface.co/ThakiCloud/SKILLRET-Reranker-0.6B) Released by ThakiCloud through its official repository on Hugging Face. [Read the license](https://www.apache.org/licenses/LICENSE-2.0). ## Built From - Derived from [Qwen/Qwen3-Reranker-0.6B](https://savrn.com/models/qwen3-reranker-0-6b) - Described by arXiv:2605.05726 - Trained on (disclosed) ThakiCloud/SKILLRET ## Memory Requirements | Precision | Weights in memory | | --- | --- | | As published | 1.2 GB | | 16-bit | 1.2 GB | | 8-bit | 0.6 GB | | 4-bit | 0.3 GB | Weights only, from the published parameter count; the key-value cache and runtime add to this. ## Questions About SKILLRET-Reranker-0.6B ### How much GPU memory does SKILLRET-Reranker-0.6B need? About 1.4 GB at 16-bit and 0.4 GB at 4-bit: the weights (596M parameters) plus a working margin. A long context needs more. ### What is the cheapest GPU to run SKILLRET-Reranker-0.6B on? At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists. ### Can I use SKILLRET-Reranker-0.6B commercially? Yes. SKILLRET-Reranker-0.6B is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors. ### What is SKILLRET-Reranker-0.6B's context length? 40,960 tokens, from the maximum position embeddings in its published configuration. ## Similar Models Model · Text ranking ### [Qwen3-Reranker-0.6B](https://savrn.com/models/qwen3-reranker-0-6b) [Qwen](https://savrn.com/model-publishers/qwen) The Qwen3 Embedding model series is the latest proprietary model of the Qwen family, specifically designed for text embedding and ranking tasks. Building upon the dense foundational models of the Qwen3 series, it provides a comprehensive range of text embeddings and reranking models in various sizes (0.6B, 4B, and 8B). This series inherits the exceptional multilingual capabilities, long-text understanding, and reasoning skills of its foundational model. The Qwen3 Embedding series represents significant advancements in multiple text embedding and ranking tasks, including text retrieval, code retrieval, text classification, text clustering, and bitext mining. Exceptional Versatility: The… Open weights apache-2.0 596M parameters 40,960 tokens transformers [View model](https://savrn.com/models/qwen3-reranker-0-6b) Model · Text ranking ### [SKILLRET-Reranker-0.6B](https://savrn.com/models/skillret-reranker-0-6b) [Anonymous Authors](https://savrn.com/model-publishers/anonymous-ed-benchmark) This is a reranker fine-tuned for AI agent skill retrieval. Given a natural-language user request and a candidate agent skill, it scores how relevant and useful the skill is for the request. It is designed as the second stage after a first-stage retriever such as SkillRet-Embedding-0.6B or SkillRet-Embedding-8B. The model is fine-tuned from Qwen/Qwen3-Reranker-0.6B on the SkillRet benchmark training split with binary cross-entropy on the yes/no token probability. It keeps the scoring interface of Qwen3-Reranker. Each skill document is body, the same representation used by the SkillRet embedding models. Evaluated on the SkillRet benchmark evaluation split (4,392 queries, 6,006 skills). The… Open weights apache-2.0 596M parameters 40,960 tokens transformers [View model](https://savrn.com/models/skillret-reranker-0-6b) Model · Text ranking ### [jina-reranker-v3](https://savrn.com/models/jina-reranker-v3) [Jina AI](https://savrn.com/model-publishers/jinaai) jina-reranker-v3 is a 0.6B parameter multilingual document reranker with a novel last but not late interaction architecture. Unlike ColBERT's separate encoding with multi-vector matching, this model performs causal self-attention between query and documents within the same context window, extracting contextual embeddings from the last token of each document. Built on Qwen3-0.6B with 28 transformer layers and a lightweight MLP projector (1024→512→256), it processes up to 64 documents simultaneously within 131K token context. The model achieves state-of-the-art BEIR performance with 61.94 nDCG@10 while being 10× smaller than generative listwise rerankers. Use transformers for local inference… Open weights cc-by-nc-4.0 597M parameters 131,072 tokens transformers [View model](https://savrn.com/models/jina-reranker-v3) Model · Text ranking ### [KaLM-Reranker-V1-Nano-R2](https://savrn.com/models/kalm-reranker-v1-nano-r2) [KaLM-Embedding](https://savrn.com/model-publishers/kalm-embedding) KaLM-Reranker-V1-R2 is a substantially improved checkpoint release of KaLM-Reranker-V1. It keeps the same fast-but-not-late-interaction (FBNL) architecture and inference interface as the original release, while introducing a stronger multi-stage training recipe for three practical goals: 1. Stronger compression robustness. R2 supports Matryoshka embedding pooling (MEP) from 1x to 128x, extending the maximum validated compression ratio from 32x to 128x. Even at 128x compression, all three model sizes retain at least 92% of their average nDCG@10 at 2x compression on both BEIR and MIRACL. 2. Adjustable test-time compute scaling. A single KaLM-Reranker-V1-R2 checkpoint can trade compute for… Open weights apache-2.0 786M parameters sentence-transformers [View model](https://savrn.com/models/kalm-reranker-v1-nano-r2) Model · Text ranking ### [jina-reranker-v2-base-multilingual](https://savrn.com/models/jina-reranker-v2-base-multilingual) [Jina AI](https://savrn.com/model-publishers/jinaai) The Jina Reranker v2 (jina-reranker-v2-base-multilingual) is a transformer-based model that has been fine-tuned for text reranking task, which is a crucial component in many information retrieval systems. It is a cross-encoder model that takes a query and a document pair as input and outputs a score indicating the relevance of the document to the query. The model is trained on a large dataset of query-document pairs and is capable of reranking documents in multiple languages with high accuracy. Compared with the state-of-the-art reranker models, including the previous released jina-reranker-v1-base-en, the Jina Reranker v2 model has demonstrated competitiveness across a series of benchmarks… Open weights cc-by-nc-4.0 278M parameters 1,026 tokens transformers [View model](https://savrn.com/models/jina-reranker-v2-base-multilingual) Model · Text ranking ### [KielEmbed-Rerank](https://savrn.com/models/kielembed-rerank) [Tech](https://savrn.com/model-publishers/kiel2) This is a Cross Encoder model finetuned from BAAI/bge-reranker-base using the sentence-transformers library. It computes scores for pairs of texts, which can be used for text reranking and semantic search. First install the Sentence Transformers library: Then you can load this model and run inference. Approximate statistics based on the first 100 samples: - perdevicetrainbatchsize: 4 - numtrainepochs: 1 - learningrate: 2e-05 - warmupsteps: 0.1 - gradientaccumulationsteps: 8 - fp16: True - perdevicetrainbatchsize: 4 - numtrainepochs: 1 - maxsteps: -1 - learningrate: 2e-05 - lrschedulertype: linear - lrschedulerkwargs: None - warmupsteps: 0.1 - optim: adamwtorchfused - optimargs: None… Open weights 278M parameters 514 tokens sentence-transformers [View model](https://savrn.com/models/kielembed-rerank) ## ThakiCloud [All models and datasets](https://savrn.com/model-publishers/thakicloud) ## Versions - [ce09c883af78](https://savrn.com/models/skillret-reranker-0-6b-2/versions/ce09c883af78) · current 2026-09-24 ## Explore More - [All text ranking models](https://savrn.com/models/tasks/text-ranking) - [All models under apache-2.0](https://savrn.com/models/licenses/apache-2-0) - [Model comparisons](https://savrn.com/models/comparisons) - [The model directory](https://savrn.com/models) - [Open model prices by host](https://savrn.com/ai-index/pricing/open-models) ## Source - Repository metadata, read 2026-09-24. - [Hugging Face record](https://huggingface.co/ThakiCloud/SKILLRET-Reranker-0.6B) - [How the hub is built](https://savrn.com/model-hub/methodology)