Sakana AI vs StepFun

Neither company has publicly disclosed a valuation, making a definitive size comparison difficult. Sakana AI employs 100-500 people, while StepFun has 100-500 employees.

StepFun has raised more in total funding at $200M, compared to Sakana AI's $200M — a gap of N/A.

Sakana AI holds the higher Awaira Score at 65/100, compared to StepFun's 58/100. The Awaira Score is a composite metric factoring in valuation, funding, stage, team size, and market presence — a 7-point gap that reflects meaningful differences in scale or traction.

Sakana AI was founded by David Ha in 2023. StepFun was founded by Jiang Daxin in 2023. Visit each company's profile on Awaira for a full founder biography.

Sakana AI: Sakana AI develops foundation models using nature-inspired evolutionary and collective intelligence approaches to AI model development, exploring alternatives to the dominant paradigm of scaling single large models toward architectures that combine multiple smaller models or that use evolutionary algorithms to discover model configurations. The Tokyo company was founded by David Ha, former Research Director at Google DeepMind, bringing frontier AI research leadership and international AI community connections to Japan AI startup ecosystem.\n\nThe company raised approximately $200 million in funding from investors including New Enterprise Associates, Lux Capital, and Khosla Ventures, reflecting global investor confidence in the founding team research pedigree. Sakana AI has published research on evolutionary model merge approaches that generate new model capabilities by combining weights from existing trained models without additional training, attracting significant attention from the global AI research community for the technical novelty and practical efficiency of the approach.\n\nSakana AI occupies a distinctive position in the Japanese AI ecosystem as a research-first foundation model lab with global ambitions, combining international founding team credentials with Japan as a base for building AI capabilities outside the US-China AI race dynamic. The company research on alternative model development approaches positions it to contribute fundamental methodological innovations that could influence the broader AI field rather than competing purely on compute scale with OpenAI and Anthropic. StepFun: StepFun develops multimodal foundation models for Chinese and English language applications, producing the Step series of large language and vision-language models that target enterprise deployment through API and private cloud licensing. The Beijing company was founded by Jiang Daxin, a former senior executive at Baidu, bringing search and large-scale AI infrastructure experience to foundation model development in the rapidly competitive Chinese LLM startup market.\n\nThe company raised approximately $200 million in early funding from Chinese venture capital firms including Hillhouse Capital and other technology-focused investors backing the wave of Chinese foundation model startups that emerged following the launch of ChatGPT. StepFun models are positioned for enterprise customers in finance, legal, and customer service sectors requiring high-quality Chinese language reasoning and instruction following capabilities.\n\nStepFun competes in a crowded Chinese foundation model market alongside Moonshot AI, MiniMax, Zhipu AI, and Baichuan, as well as the AI divisions of Alibaba, Baidu, and ByteDance that have substantially greater compute resources and distribution. The Chinese foundation model market has seen aggressive fundraising and model release activity as the government signals support for domestic AI capability development. Differentiation among Chinese LLM startups is increasingly difficult as model benchmark performance converges, shifting competition toward vertical domain adaptation, inference cost, and enterprise integration services.

Both Sakana AI and StepFun were founded in the same year — 2023. Despite sharing a founding year, they may have launched at different times within that year, which can matter in fast-moving AI markets.

Both Sakana AI and StepFun report similar employee counts of approximately 100-500. Team size is often a proxy for operational scale, though lean AI companies can punch well above their headcount.

Yes, Sakana AI and StepFun are direct competitors — both operate in the Foundation Models space and likely target overlapping customer segments. This comparison is especially relevant for buyers evaluating both platforms.