allworld.mediaNews from all over the world
StepFun's Step 5 Preview available for free in Tencent WorkBuddy until October 25
Read more
Pandaily
pandaily.com

StepFun's Step 5 Preview available for free in Tencent WorkBuddy until October 25

StepFun announced on October 12 that its Step 5 Preview model is now integrated into WorkBuddy—Tencent's AI-powered office agent. From October 12 to October 25, users can utilize Step 5 Preview for free within WorkBuddy, which StepFun termed an exclusive agreement, according to IT Home.

The announcement concerned only the access and usage period. StepFun did not provide information on how the pricing will look in WorkBuddy after the two-week period ends, or whether this model will become a permanent option in the product.

For WorkBuddy users, this expands the selection of models within the agent, which is already capable of performing various office tasks such as document handling, spreadsheet management, and scheduling, on both desktop and mobile devices. For StepFun itself, this means its newest flagship product is accessible to a broad base of office users, not just developers accessing the API.

The Step 5 Preview model is not new; it was introduced by StepFun on September 20 as a flagship focused on AI-assisted coding, software development, professional knowledge work, and finance, with a special emphasis on real-world agent tasks. The model utilizes a sparse mixture-of-experts architecture with a total of 600 billion parameters and 27 billion active parameters, supports a context window of 1 million tokens, and initially accepts both text and visual input.

StepFun stated that it will publish the full open-source model weights on October 15, three days after the promotion begins in WorkBuddy. This allows WorkBuddy users to test the deployed model during the same period when developers can download and run it independently. Pandaily previously covered the launch and architecture of this model.

WorkBuddy already allows users to choose between models from various external laboratories. Integration with these external labs enriches the agent's menu for performing diverse tasks, while model creators gain a distribution channel within one of China's most frequently used AI office tools.

Similar stories

StepFun releases preliminary version of Step 5 model: agent with 600 billion parameters and open weights expected on October 15
Read more
pandaily.com

StepFun releases preliminary version of Step 5 model: agent with 600 billion parameters and open weights expected on October 15

StepFun introduced the preliminary version of Step 5 on September 20, 2026, as its next flagship foundational model designed for long-horizon agents, according to information from Tencent Tech, DataLearner model cards, and company materials available at stepfun.com. The official English name of the model is StepFun / Step 5.

It is important to note that although the model identifier step-5-preview is already available via the product API and the open StepFun platform, the open weights in BF16 format are scheduled only for October 15, 2026, and were not available at launch. Therefore, availability via API and open weights should be considered as separate points.

The model architecture features a Mixture of Experts (MoE) sparse design with an approximate total of 600 billion parameters. Approximately 27 billion parameters are activated per token within a 92-layer 'narrow and deep' Transformer. The choice of depth is driven by agent requirements, as longer information paths through layers contribute to improved multi-step implicit reasoning and handling of long tool outputs.

The model supports a one-million-token context window, accepting both text and image input (though video input is mentioned on third-party cards). It is oriented towards applications in AI coding, software development, financial analysis, and professional agent usage. To maintain practicality with the million-token attention, sparse GQA was applied in combination with token block merging, reducing the cost of indexer and top-k selection by approximately one eighth.

During training, special attention was paid to bit-level alignment between training and inference to ensure stable MoE routing, and load-aware scheduling, speculative decoding, and FP8 paths were utilized. StepFun claims that these methods accelerate the long-horizon Reinforcement Learning (RL) process by more than three times overall.

According to data from Artificial Analysis's composite AI index, it scores 44 points. The company places this rating among leading models focused on open access. Aggregator cards also list API prices: about $1.00 per input token and $2.70 per output token per million tokens, but comparisons of metrics and cost should be viewed as statements from third parties or providers.

Until October 15, Step 5 Preview should primarily be regarded as a functional API for long-horizon agents, having only a planned commitment for open weights, rather than a release of weights on Hugging Face or a drop-in replacement for Meituan LongCat, MiniMax Code Flash, or NaiveAI OSS MoE.

Popular