StepFun
PulseAugur coverage of StepFun — every cluster mentioning StepFun across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
StepFun's STEPX Neo AI phone integrates deeply with Chinese super apps
StepFun has launched the STEPX Neo AI phone in China, featuring deep integrations with popular super apps such as Alipay, Meituan, and DiDi. This device can autonomously perform complex, multi-step tasks across these pl…
-
StepFun launches STEPX Neo AI agent phone with deep app integration
StepFun has launched the STEPX Neo, an AI agent-native smartphone designed for autonomous task execution across multiple applications. The device integrates over ten major platforms, including Alipay, Meituan, DiDi, and…
-
ZTE Nubia and StepFun launch competing AI agent smartphones
The race for AI agent smartphones is intensifying with competing launches from ZTE Nubia and StepFun. ZTE Nubia has unveiled what it claims is the world's first AI agent phone at the World Artificial Intelligence Confer…
-
Huawei's Atlas 950 SuperPoD and first AI agent phone to debut at China AI summit
Huawei is set to unveil its Atlas 950 SuperPoD computing cluster at the upcoming World Artificial Intelligence Conference (WAIC) in Shanghai. This cluster, previously revealed last year, is designed for AI and machine l…
-
Multi-tier MoE caching discussed as future of LLM inference
A discussion on Reddit explores the concept of multi-tier Mixture of Experts (MoE) caching as a potential future direction for MoE model inference. The idea involves strategically distributing model experts across CPU a…
-
StepFun releases Step 3.7 Flash with vision and auto-escalation
StepFun has released Step 3.7 Flash, an upgraded version of its 3.5 Flash model, featuring a new vision encoder and an automatic "Advisor Mode" that escalates complex tasks to larger models. This update aims to improve …
-
Stepfun's Step 3.7 Flash matches Claude's coding at 1/9 cost
Chinese AI startup Stepfun has released its Step 3.7 Flash model, which reportedly achieves 97% of Claude's coding capabilities at one-ninth the cost. This new model offers a speed of 416 tokens per second, making it a …
-
Chinese AI Firm StepFun Prepares for Hong Kong IPO
Chinese AI startup StepFun is reportedly preparing to file for an initial public offering on the Hong Kong Stock Exchange. The company, which focuses on AI technologies, aims to raise capital through this public listing…
-
Step 3.7 Flash leads benchmarks in speed, cost, and performance
StepFun's new model, Step 3.7 Flash, has achieved top rankings on the Artificial Analysis (AA) benchmark, excelling in speed, cost-efficiency, and end-to-end performance. The model demonstrates impressive output speeds …
-
StepFun's Step 3.7 Flash leads AI output speed benchmarks
StepFun's new open-source model, Step 3.7 Flash, has achieved the top position on Artificial Analysis's Output Speed leaderboard, processing 409 tokens per second. The model also leads in other key metrics like end-to-e…
-
AliExpress 618 Sale Sees Near 40% Brand GMV Penetration
AliExpress's summer 618 promotion, which began on June 1st, has seen a significant surge in brand Gross Merchandise Volume (GMV) penetration, reaching nearly 40% on its opening day. This event has facilitated a concentr…
-
Open Source LLMs See Underwhelming May 2026 Release Pace
The open-source large language model landscape saw a relatively quiet May 2026, following a strong April. Despite the release of models like Ring, Command, and StepFun, the month was considered underwhelming by enthusia…
-
StepFun's Flash Model Nears Claude Opus Coding Performance at Fraction of Cost
StepFun has released its Step 3.7 Flash model, which reportedly achieves 97% of the coding performance of Anthropic's Claude Opus 4.6. This new model is significantly more cost-efficient, operating at one-ninth the cost…
-
StepAudio 2.5 beats OpenAI and Google in emotion tests
StepAudio 2.5, a new model from Chinese startup StepFun, has reportedly surpassed OpenAI and Google systems in emotion interpretation tests. The model was trained using 10,000 base patterns.
-
Model catalog sees new additions, price changes, and removals
New models have been added to the model catalog, including StepFun's Step 3.7 Flash, which offers large-context generation at a moderate cost. Anthropic has released two new Claude Opus 4.8 variants: a "Fast" version fo…
-
StepFun releases 198B MoE vision-language model for agents
StepFun has released Step 3.7 Flash, a 198 billion parameter Mixture-of-Experts vision-language model designed for coding agents and search workflows. This new model features native multimodal understanding, improved to…
-
China tightens grip on AI talent with travel bans and investment curbs
China is implementing stricter measures to retain its top AI talent, including travel restrictions and government approval requirements for researchers and startup founders. This move is seen as a strategic effort to sa…
-
StepFun launches StepAudio 2.5 Realtime voice AI
Chinese AI lab StepFun has released StepAudio 2.5 Realtime, a new end-to-end live voice model. This system handles both audio input and output, incorporating features for persona control and paralinguistic understanding.
-
StepFun launches StepAudio 2.5 with real-time voice and persona consistency
StepFun has released StepAudio 2.5 Realtime, an end-to-end speech large language model capable of real-time, customizable persona interactions. The model integrates speech understanding and generation, utilizing a milli…
-
Stepfun AI releases 198B parameter multimodal MoE model
Stepfun AI has released Step 3.7 Flash, a 198-billion parameter sparse Mixture-of-Experts (MoE) vision-language model. This model is optimized for agentic workflows, coding, and multimodal tasks, activating approximatel…