Why high: an exact fingerprint shared with 13 models that declare the same family, without a second signal of a different kind yet.
StepFun: Step 3.7 Flash is StepFun's model. Its measured tokenizer sits with the DeepSeek group, which is evidence that it builds on the DeepSeek tokenizer. Tokenizer reuse is normal engineering, open vocabularies travel between labs, and this says nothing against StepFun's authorship of the model itself.
As with Step 3.5 Flash: StepFun's own Step-3 repository ships a tokenizer behaviourally identical to DeepSeek-V3's over all 50 strings, and this endpoint counts the same way. Its place in the DeepSeek signature group is genuine lineage reuse by StepFun.
| We checked | What we found | Where it came from |
|---|---|---|
| How it counts tokens | Counts every one of the 50 test strings exactly like 13 other models | tk_73c4f0dc |
| How its API is set up | A combination of settings no other listing in the catalogue uses | api_25b530bb |
| How much it can read and write | Reads up to 262,144 tokens at once, writes up to 230,400 | its own listing |
| Whether it thinks before answering | Always reasons, and cannot be turned off, at high, medium, low, medium by default | its own listing |
| Who served our requests | StepFun | we measured it, 5 October 2026 |
| Measured twice | The 5 October 2026 reading agreed with the 21 August 2026 reading on every one of the 50 strings clean in both: the tokenizer behind this name held. | 50/50 agree |
Click a row to open the full comparison.
We sent StepFun: Step 3.7 Flash fifty short pieces of text and recorded what each one cost it in tokens. The bars below are those costs. Two models built on the same tokenizer produce the same bars; a model built on a different one produces a different set, which is what makes this a fingerprint.
Overhead subtracted: 12 prompt tokens.
Step 3.7 Flash is StepFun's latest high-efficiency multimodal Mixture-of-Experts model. It pairs a 196B-parameter language backbone with a vision encoder for native image and video understanding, activating roughly 11B parameters...
Supported parameters
frequency_penaltyinclude_reasoninglogprobsmax_tokenspresence_penaltyreasoningreasoning_effortrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p
defaults: {"temperature":null,"top_p":null,"top_k":null,"frequency_penalty":null,"presence_penalty":null,"repetition_penalty":null}
