Train their own language models, have raised funding, work on multi-step reasoning

Moreh trains their own language models, has raised $22M in Series B funding, and their MoAI platform is used for multi-step reasoning in large AI models.

Neuron7 trains their own language models, has raised funding in the last year, and works on multi-step reasoning technologies.

Adept matches the query because they develop and train their own language models (Fuyu-8B and Persimmon-8B), have raised $350M in Series B funding in the last year, and are involved in research and development of multi-step reasoning technologies as evidenced by their work on modeling abstract goals and understanding software.

Atomic AI develops and trains their own language models, has raised $35 million in Series A financing, and is involved in research and development of multi-step reasoning technologies through their AI-driven RNA Structure Platform.

AssemblyAI develops and trains their own language models, as evidenced by their Conformer-2 model and next-gen Universal model. They have raised over $65M in funding, with 90% of the total funds raised in the last 22 months. They are involved in multi-step reasoning technologies, as indicated by their work on Audio Intelligence models and LeMUR for text generation tasks over recognized speech.

Harvey develops and trains their own language models tailored for legal tasks, has raised over $100 million in funding, and leverages large language models for complex legal tasks which likely involve multi-step reasoning.

Automat develops and trains their own language models as they utilize large language models and computer vision for automation. They have raised $3.75M in a seed round, which indicates they have raised funding in the last year. Additionally, the company is involved in research and development of multi-step reasoning technologies, as evidenced by their use of generative AI, natural language processing, and their founders' background in conversational interfaces and LLM research at Google.

OpenAI develops and trains their own language models, has raised significant funding in recent years, and is involved in research and development of multi-step reasoning technologies.

Imbue (formerly Generally Intelligent) matches the query because: 1) The company develops and trains their own language models, as evidenced by their work on AI systems that can reason and code, and the development of CARBS, a cost-aware hyperparameter optimizer. 2) The company has raised $200M in funding recently, indicating they have raised funding in the last year. 3) The company is involved in research and development of multi-step reasoning technologies, as shown by their focus on AI systems that can reason and code, and their contributions to research on LLM ethical decision-making and efficient training.