What are the hardware requirements for Qwen3 Swallow 32B RL V0.2?

Qwen3 Swallow 32B RL V0.2 requires approximately 65.1 GB of VRAM and supports a context window of 40K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Qwen3 Swallow 32B RL V0.2 and how large is it?

Qwen3 Swallow 32B RL V0.2 is developed by tokyotech-llm, a model with 32b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Qwen3 Swallow 32B RL V0.2?

Qwen3 Swallow 32B RL V0.2 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Qwen3 Swallow 32B RL V0.2 by tokyotech-llm — VRAM 65.1GB, 40K context

Name: Qwen3 Swallow 32B RL V0.2
Author: tokyotech-llm

Qwen3 Swallow 32B RL V0.2 is an open-source language model by tokyotech-llm. Features: 32b LLM, VRAM: 65.1GB, Context: 40K, License: apache-2.0, LLM Explorer Score: 0.28.

Arxiv:2404.17733 Arxiv:2412.02595 Arxiv:2505.09388 Base model:finetune:tokyotech-... Base model:tokyotech-llm/qwen3... Conversational Dataset:tokyotech-llm/lmsys-ch... Dataset:tokyotech-llm/swallow-... Dataset:tokyotech-llm/swallow-... Dataset:tokyotech-llm/swallow-... En Endpoints compatible Ja Qwen3 Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/tokyotech-llm/Qwen3-Swallow-32B-RL-v0.2

Qwen3 Swallow 32B RL V0.2 Parameters and Internals

LLM Name	Qwen3 Swallow 32B RL V0.2
Repository 🤗	https://huggingface.co/tokyotech-llm/Qwen3-Swallow-32B-RL-v0.2
Base Model(s)	tokyotech-llm/Qwen3-Swallow-32B-SFT-v0.2 tokyotech-llm/Qwen3-Swallow-32B-SFT-v0.2
Model Size	32b
Required VRAM	65.1 GB
Updated	2026-07-18
Maintainer	tokyotech-llm
Model Type	qwen3
Model Files	5.3 GB: 0-of-13 5.3 GB: 1-of-13 5.2 GB: 2-of-13 5.2 GB: 3-of-13 5.3 GB: 4-of-13 5.3 GB: 5-of-13 5.2 GB: 6-of-13 5.2 GB: 7-of-13 5.2 GB: 8-of-13 5.2 GB: 9-of-13 5.2 GB: 10-of-13 5.2 GB: 11-of-13 2.3 GB: 12-of-13
Supported Languages	en ja
Model Architecture	Qwen3ForCausalLM
License	apache-2.0
Context Length	40960
Model Max Length	40960
Transformers Version	4.57.3
Tokenizer Class	Qwen2Tokenizer
Padding Token	<\|endoftext\|>
Vocabulary Size	151936
Errors	replace

Rank the Qwen3 Swallow 32B RL V0.2 capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Qwen3 Swallow 32B RL V0.2

Best Alternatives	Context / RAM	Downloads	Likes
...ll Qwen3 32B Preview7 QAT 200K	195K / 65.8 GB	10	3
...ll Qwen3 32B Preview6 QAT 200K	195K / 65.8 GB	6	2
...ll Qwen3 32B Preview4 QAT 200K	195K / 65.8 GB	13	2
Qwen3 32B AWorld	128K / 65.8 GB	37	15
Qwen3 32B	40K / 65.8 GB	5411	15
Qwen3 32B	40K / 65.6 GB	6155624	715
OpenThinkerAgent 32B	40K / 65.8 GB	9022	9
EviNurse 32B	40K / 65.5 GB	107	0
Qwen3 Swallow 32B SFT V0.2	40K / 65.6 GB	24411	0
...gy GPT Regulatorio 32B Safe V2	40K / 131.6 GB	205	0

Note: green Score (e.g. "73.2") means that the model is better than tokyotech-llm/Qwen3-Swallow-32B-RL-v0.2.