Step-3.7-Flash-FP8

Checkpointstepfun-ai/Step-3.7-Flash-FP8Model card indexedModel card

Support summary

Step-3.7-Flash-FP8 is a checkpoint published by StepFun, built on the Step3p7ForConditionalGeneration architecture. Official recipes or cookbooks cover vLLM, without pinning a stable release.

Documented quantization: FP8
Documented features: Multi-token prediction, 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: StepFun
Task: image text to text
License: apache-2.0
Library: transformers
Architecture: Step3p7ForConditionalGeneration
Weight tensor elements: 201.4B
Tensor types: BF16, F32, F8_E4M3
Quantization: fp8
Configured context: 262,144 tokens
Languages: en
Weight access: Public
Repository created: 2026-05-23

Source 1 Source 2 Source 3Reviewed 2026-09-05

Model notes and sources

Engine support

3 engines

Engine support for Step-3.7-Flash-FP8
Engine	Release	Evidence scope	Last known supported
vLLM	Version not pinnedRecipe / cookbook	Documented recipe Official recipe Source: vllm · models/stepfun-ai/Step-3.7-Flash.yamlReviewed 2026-09-05	No supported release indexed
Features, dependencies and evidence for vLLM 0.28.0
SGLang	`0.5.19`Newest indexed	No matching evidence curated	No supported release indexed
Features, dependencies and evidence for SGLang 0.5.19
TensorRT-LLM	`1.2.1`Newest indexed	No matching evidence curated	No supported release indexed
Features, dependencies and evidence for TensorRT-LLM 1.2.1

Engine support

Recipes and cookbooks