StepFun: Step 3.5 Flash

Step 3.5 Flash is a text model from StepFun, listed on api14 under the model id stepfun/step-3.5-flash. StepFun describes it as their most capable open-source foundation model, built on a sparse Mixture of Experts architecture that activates only 11B of its 196B parameters per token. It sits in the text category, so it is aimed at language work rather than image, audio or other modalities.

The model takes text in and returns text out. It supports a context window of 262144 tokens, which is large enough to hold long documents, extended conversations or sizeable code files in a single request, and it can produce up to 65536 tokens in its output. Among its supported parameters are reasoning and include_reasoning, so the model can expose a reasoning mode alongside its normal responses. It also accepts tools and tool_choice, which means it can be given function definitions and asked to call them.

Those capabilities point to a few practical uses. The long context suits summarising or querying long source material, such as technical documentation, contracts or multi-file codebases, without splitting them into chunks first. The reasoning parameter fits problems that benefit from step-by-step working, for example multi-step analysis or debugging. Tool calling makes it a candidate for agent-style workflows where the model decides which function to invoke, and the 65536-token output ceiling helps when the answer itself is long, such as a generated report or a large refactor.

You call it through api14's OpenAI-compatible API using the model id stepfun/step-3.5-flash, so existing OpenAI client libraries and request shapes work with only the model name changed. The request parameters it supports are frequency_penalty, include_reasoning, max_tokens, reasoning, temperature, tool_choice, tools, top_k and top_p. In practice you set temperature, top_p or top_k to steer sampling, use max_tokens to cap the response, pass tools and tool_choice for function calling, and toggle reasoning or include_reasoning when you want the model's reasoning behaviour in the response.

The main constraint to keep in mind is that this is a text-only model: it does not accept image or audio input, so a multimodal task needs a different model. If your workload is purely text and fits within the 262144-token context and 65536-token output limits, Step 3.5 Flash is a reasonable choice, particularly when you want reasoning and tool calling in the same model.