DeepSeek-V4-Flash
Support summary
DeepSeek-V4-Flash is a checkpoint published by DeepSeek, built on the DeepseekV4ForCausalLM architecture. vLLM 0.28.0, the newest indexed vLLM release, documents support. SGLang last documented support in 0.5.18; its newest indexed release does not list it.
vLLM documents support in 14 of 20 indexed releases, from 0.20.0 to 0.28.0.
- Documented hardware
- NVIDIA / CUDA and AMD / ROCm
- Documented quantization
- fp4 and FP8
- Documented features
- Multi-token prediction, Reasoning parsers, Speculative decoding, Tensor parallelism, and Tool calling
- Configured context
- 1,048,576 tokens
- License
- mit
Hardware, quantization and features count only statements about this checkpoint; engine-wide capabilities are listed with each release.
Hugging Face model card
- Publisher
- DeepSeek
- Task
- text generation
- License
- mit
- Library
- transformers
- Architecture
- DeepseekV4ForCausalLM
- Weight tensor elements
- 290.9B
- Tensor types
- BF16, I64, F32, F8_E4M3, I8
- Quantization
- fp8
- Configured context
- 1,048,576 tokens
- Weight access
- Public
- Repository created
- 2026-04-22
Model notes and sources
- vLLM v0.28.0 · docs/models/supported_models.md · 14 indexed revisions, v0.20.0 to v0.28.0
- deepseek-ai/DeepSeek-V4-Flash · publisher model card
- deepseek-ai/DeepSeek-V4-Flash · Hugging Face metadata
- deepseek-ai/DeepSeek-V4-Flash · publisher configuration
- vllm · models/deepseek-ai/DeepSeek-V4-Flash.yaml
- sglang · docs/src/snippets/configs/deepseek-ai/deepseek-v4.jsx
- sglang · docs/cookbook/autoregressive/DeepSeek/DeepSeek-V4.mdx
- Reviewed
Engine support
3 engines
| Engine | Release | Checkpoint support | Evidence scope | Last known supported |
|---|---|---|---|---|
| vLLM | 0.28.0Newest indexed | Checkpoint evidenceSource: vLLM v0.28.0 · docs/models/supported_models.mdReviewed | 0.28.0This release | |
Features, dependencies and evidence for vLLM 0.28.0 | ||||
| SGLang | Varies by configurationRecipe / cookbook | Documented recipeSource 1: sglang · docs/cookbook/autoregressive/DeepSeek/DeepSeek-V4.mdxSource 2: sglang · docs/src/snippets/_deployment.jsxSource 3: sglang · docs/src/snippets/configs/deepseek-ai/deepseek-v4.jsxSource 4: sglang · docs/src/snippets/configs/deepseek-ai/deepseek-v4-benchmarks.jsxSource 5: sglang · docs/src/snippets/_playground.jsxReviewed | View 0.5.18 | |
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 | ||||