HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks

Qwen3.5-27B

Qwen/Qwen3.5-27B

kind
original
quant
bf16 (16-bit)
family
qwen
modality
image-text-to-text
root model
Qwen/Qwen3.5-27B
provider
Qwen
license
apache-2.0
language
—
context
262,144
library
transformers
vRAM (GB)
—
size on disk
—
downloads
2,599,019
likes
1,018
created
—
last modified
2026-04-24T02:54:49+00:00 (2026)
huggingface
https://huggingface.co/Qwen/Qwen3.5-27B
tags
transformerssafetensorsqwen3_5image-text-to-textconversationallicense:apache-2.0eval-resultsendpoints_compatibledeploy:azureregion:usdeploy:sagemaker

Benchmark scores (0)

no scores extracted for this variant.

Discussions (51)

titleauthorstatuscommentstype
Listed on OpenModelMapduola15open0disc
Qwen3.5-27B brain atlas activation census, OV-circuits, and capability fence (64layers)juiceb0xc0deopen0disc
Listed on OpenModelMapduola15open0disc
Listed on OpenModelMapduola15open0disc
Add EvalEval community eval resultsEvalEvalBotopen0disc
Add MMMU-Pro evaluation resultSaylorTwiftopen0disc
qwen3.5-27b + openclawRZZRopen0disc
fix chat template to avoid empty historical `<think>` blockslatent-variableopen0disc
Qwen 3.5 27b fine tuning multi-gpu example lora sftEthermichopen0disc
Heavy Chinese-English mixingariesaaopen0disc
65454yehuashanguidraft0disc
what the best engine to run this model.Shimon324open0disc
qwen3.5-27b,快速模式下,发票的年份识别不确认,2026识别成2025,在思考模式下才正常htl258open0disc
Add ScreenSpot-Pro evaluation resultmerveopen0disc
what's the matrics of livecode bench v6 when enable think offHoldenShenopen0disc
repeat_penalty not repetitionpants44open0disc
Add MathArena evaluation result for hmmt/hmmt_feb_2026JasperDekoninckopen0disc
Add MathArena evaluation result for aime/aime_2026JasperDekoninckopen0disc
`.cw` is output at the end of the reasoning contentowaoopen0disc
Chinese-English Typography Issuesdouglarekopen0disc
Looping foreverkil3ropen0disc
[transformers-server] Continuous batching does not support Qwen-3.5 serieszhang-zhi-chunopen0disc
Language Request: Garhwaliakshitgaur2005open0disc
What is the chat template for Ollama ?ngbruceopen0disc
tutorbububabyopen0disc
Fine tune with lorasaireddyopen0disc
Please don't forget to release the base model of Qwen3.5-27BLinksomeopen0disc
pure transformers inference code is neededmaltosefloweropen0disc
When?AxionLab-officialopen0disc
as a starterkit17open0disc
Upload 3 filesshamamanaeem03open0disc
model overthinkingcse2011open0disc
vLLM Version conflict issueinnosynthopen0disc
有没有BFCLV4和tau2的细节指标HuanchangLiuopen0disc
GGUFblankregclosed0disc
Thinking Blocks ErrorTronixATopen0disc
Plan on releasing the 4bits quantized versionomarsouopen0disc
Qwen are is the 8b coming out?Crowneliusopen0disc
vllm serving issuegrozatechopen0disc
Russian language support, bad grammar!alexcardoopen0disc
Add evaluation resultsSaylorTwiftopen0disc
Qwen3.5-27B-FP8Johnassonclosed0disc
Availability of the Qwen3.5-27B Base modelJellyfish042open0disc
Value error, Model architectures ['Qwen3_5ForConditionalGeneration'] are not supported for now. Transformers version 5.3.0.dev0rameshchopen0disc
Update README.mdbzhengmerged0disc
Installation Video and Testing - Step by Stepfahdmirzacopen0disc
amazing modelTugay31open0disc
is HLE w/ CoT a typo?xiaoqianWXopen0disc
Samplersggnoyopen0disc
DENSEox-oxopen0disc
请千万千万别忘了把 Qwen Image 2.0 也开源——这对我们本地用户来说会是个巨大的改变 :-)Hanswalteropen0disc
Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.