Sber presents "GigaAgent": an autonomous AI agent with self-development and continuous learning capabilities
Sber has officially launched "GigaAgent" — a universal autonomous assistant designed to handle everyday and work tasks. This step marks a new stage in the race for leadership in the field of AI agents, where competition is intensifying every day.
The market is rapidly moving away from simple chatbots toward full-fledged digital partners capable of planning and executing multi-step processes. "GigaAgent" is the Russian answer to foreign solutions like OpenClaw, Hermes Agent, and NemoClaw. The product stands out from its analogs thanks to three key characteristics: autonomy, self-improvement, and continuity.
Assistant for routine tasks
Unlike conversational chats, the new agent does not require control at every step. It manages the calendar, works with documents, handles correspondence, searches for information, and prepares reports and presentations. Moreover, it can operate in the background, take initiative, ask clarifying questions, and even raise reasoned objections when data is insufficient.
The key difference is the ability for self-improvement. "GigaAgent" can read and rewrite its own source code, recording each change as a commit in a built-in repository. This allows it not only to perform tasks but also to refine its algorithms. Additionally, it remembers past experience and context, resuming work after a restart from where it left off.
Technology and pricing
The development is based on the self-improving agent project Ouroboros from the AIRI research institute. On the Terminal Bench 2.1 benchmark for Anthropic's Opus 5 model, the agent demonstrates an impressive result of 86.97%.
The service can be launched through the Agents Space platform from Cloud.ru. New users are offered a starting grant of 4,000 rubles for free testing of the capabilities. Further usage is billed depending on the use of language models and infrastructure, with the final amount directly depending on the volume and frequency of requests.
Investments in the project, according to expert estimates, amounted to approximately 1–3 billion rubles over a year and a half. Assembling a team of about 100 people and developing orchestration, connectors, and a quality assessment system required significant resources, although the base model GigaChat 3 was already available under a free license.
My analysis: The launch of "GigaAgent" is not just a technological breakthrough, but a strategic move in the battle for the corporate market. However, behind the loud claims of self-improvement and autonomy, there must be measurable business metrics and clear control mechanisms. Without clear KPIs and a transparent performance evaluation system, such products risk remaining expensive experiments rather than working tools. Integration into the banking sector and the crypto industry could be the next logical step, given Sber's previously announced plans for AI banking.