Qwen
Qwen3.5-4B
Answers a written prompt using information from a picture.
Large浏览模型
可下载并运行的开放模型。选择你的内存，看看哪些放得下。
模型结果
以下是最接近且放得下 16 GB 的。
QW
ReadyA model for working with images and text together.
Large约需 12 GB 内存理解图像8.88 GBapache-2.0↓ 4M目录下载 4,037,454 次
RE
ReadyReliquaryForge
qwen3-4B-base-dapo-v4
Generates text by continuing a written prompt.
Large约需 11 GB 内存聊天与写作8.04 GBapache-2.0↓ 936.3K目录下载 936,257 次
QW
ReadyQwen
Qwen3-4B
A 4B chat model with optional reasoning and support for 100+ languages.
Large约需 11 GB 内存聊天与写作8.04 GBapache-2.0↓ 15.5M目录下载 15,512,125 次
QW
ReadyScores how closely a passage matches a search question.
Large约需 11 GB 内存Rank search results8.04 GBapache-2.0↓ 1.8M目录下载 1,755,665 次
QW
ReadyTurns text into number lists for search and meaning comparisons.
Large约需 11 GB 内存搜索文档8.04 GBapache-2.0↓ 2.4M目录下载 2,446,340 次
MI
Readymicrosoft
Phi-3.5-mini-instruct
A 3.8B instruction model for code and reasoning, with a 128K-token context window.
Large约需 10 GB 内存聊天与写作7.64 GBmit↓ 1.2M目录下载 1,162,551 次
MI
Readymicrosoft
Phi-tiny-MoE-instruct
A 3.8B English instruction model that activates 1.1B parameters per token.
Large约需 10 GB 内存聊天与写作7.51 GBmit↓ 831.7K目录下载 831,711 次
GO
ReadyCompares a picture with written labels you supply.
Large约需 10 GB 内存图像分类7.49 GBapache-2.0↓ 962.4K目录下载 962,356 次
AL
ReadyAlibaba-NLP
gte-Qwen2-1.5B-instruct
Turns text into number lists for search and meaning comparisons.
Large约需 10 GB 内存比较句子7.1 GBapache-2.0↓ 664.4K目录下载 664,441 次
DE
Readydeepseek-ai
DeepSeek-OCR-2
A model for working with images and text together.
Large约需 9 GB 内存理解图像6.78 GBapache-2.0↓ 3.2M目录下载 3,202,247 次
QW
ReadyA 3B multilingual chat model for writing and structured JSON output.
Large约需 9 GB 内存聊天与写作6.17 GBqwen-research-non-commercial↓ 5.3M目录下载 5,284,682 次
HU
ReadyHuggingFaceTB
SmolLM3-3B
A 3B chat model with optional reasoning and support for six languages.
Large约需 8 GB 内存聊天与写作6.15 GBapache-2.0↓ 663.4K目录下载 663,378 次
OR
Readyornith-ai
Ornith-1.0-9B-GGUF
A 9B reasoning model in GGUF, with tool-calling and coding examples on its model card.
Medium约需 8 GB 内存聊天与写作5.63 GBmit↓ 4.9M目录下载 4,907,682 次
QW
ReadyAn AWQ edition of a 7B multilingual chat model for writing and structured JSON output.
Medium约需 8 GB 内存聊天与写作5.57 GBapache-2.0↓ 3.5M目录下载 3,487,953 次
LE
Readylegraphista
glm-4-9B-chat-IMat-GGUF
A chat model for conversations, questions and writing.
Medium约需 7 GB 内存聊天与写作5.25 GBglm-4↓ 1M目录下载 1,013,638 次
Ready
An FP8 edition of Qwen3’s 4B direct-answer model, with native 256K-token context.
Medium约需 7 GB 内存聊天与写作5.19 GBapache-2.0↓ 873.2K目录下载 873,213 次
MU
Readymudler
locate-anything.cpp-gguf
Finds objects in a picture and marks their locations.
Medium约需 7 GB 内存Find objects in images4.72 GBnvidia-license↓ 916K目录下载 915,963 次
QW
ReadyQwen
Qwen3-ASR-1.7B
Turns recorded speech into written text.
Medium约需 7 GB 内存语音转录4.7 GBapache-2.0↓ 1.5M目录下载 1,467,010 次
GO
ReadyCompares a picture with written labels you supply.
Medium约需 6 GB 内存图像分类4.54 GBapache-2.0↓ 661.7K目录下载 661,721 次
GO
ReadyCompares a picture with written labels you supply.
Medium约需 6 GB 内存图像分类4.54 GBapache-2.0↓ 835.3K目录下载 835,289 次
QW
ReadyA model for working with images and text together.
Medium约需 6 GB 内存理解图像4.42 GBapache-2.0↓ 3.7M目录下载 3,696,917 次
QW
ReadyQwen
Qwen3-1.7B
A 1.7B chat model with optional reasoning and a native 32K-token context window.
Medium约需 6 GB 内存聊天与写作4.06 GBapache-2.0↓ 5.7M目录下载 5,748,993 次
VI
Readyvikhyatk
moondream2
A model for working with images and text together.
Medium约需 6 GB 内存理解图像3.85 GBapache-2.0↓ 1.5M目录下载 1,543,038 次