Qwen3.8-27B-FP8
Support summary
Qwen3.8-27B-FP8 is a checkpoint published by Qwen, built on the Qwen3_5ForConditionalGeneration architecture. Official recipes or cookbooks cover vLLM and SGLang, without pinning a stable release.
- Documented hardware
- Ascend NPU, NVIDIA / CUDA, and AMD / ROCm
- Documented quantization
- FP8
- Documented features
- Prefix caching, Reasoning parsers, Speculative decoding, Tensor parallelism, and Tool calling
- Configured context
- 262,144 tokens
- License
- apache-2.0
Hardware, quantization and features count only statements about this checkpoint; engine-wide capabilities are listed with each release.
Hugging Face model card
- Publisher
- Qwen
- Task
- image text to text
- License
- apache-2.0
- Library
- transformers
- Architecture
- Qwen3_5ForConditionalGeneration
- Weight tensor elements
- 27.8B
- Tensor types
- BF16, F8_E4M3
- Quantization
- fp8
- Configured context
- 262,144 tokens
- Base models
- Qwen/Qwen3.8-27B
- Weight access
- Public
- Repository created
- 2026-08-13
Publisher deployment guidance
- vLLMVersion unspecified
The model card recommends current framework versions and links a deployment recipe; it does not pin a stable release.
Source: Qwen/Qwen3.8-27B-FP8 · publisher model cardReviewed - SGLangVersion unspecified
The model card recommends current framework versions and links a deployment recipe; it does not pin a stable release.
Source: Qwen/Qwen3.8-27B-FP8 · publisher model cardReviewed
Model notes and sources
Engine support
3 engines
| Engine | Release | Checkpoint support | Evidence scope | Last known supported |
|---|---|---|---|---|
| vLLM | Version not pinnedRecipe / cookbook | Documented recipeSource: vllm · models/Qwen/Qwen3.8-27B.yamlReviewed | No supported release indexed | |
Features, dependencies and evidence for vLLM 0.28.0 | ||||
| SGLang | Version not pinnedRecipe / cookbook | Documented recipeSource 1: sglang · docs/cookbook/autoregressive/Qwen/Qwen3.8-27B.mdxSource 2: sglang · docs/src/snippets/_deployment.jsxSource 3: sglang · docs/src/snippets/configs/Qwen/qwen3.8-27b.jsxSource 4: sglang · docs/src/snippets/_qwen38_mamba_ratio_calculator.jsxSource 5: sglang · docs/src/snippets/_playground.jsxReviewed | No supported release indexed | |
Features, dependencies and evidence for SGLang 0.5.19 | ||||
| TensorRT-LLM | 1.2.1Newest indexed | No matching evidence curated | No supported release indexed | |
Features, dependencies and evidence for TensorRT-LLM 1.2.1 | ||||