PulseAugur
EN
LIVE 12:15:26
Русский(RU) T-Search: как мы обучали агента многошаговому поиску Всем привет! На связи Толя Потапов, MLE в Т-Банке. Продолжаем развивать линейку открытых моделей Gen-T. Мы

T-Bank releases T-Search agent for multi-step search

By PulseAugur Editorial · [1 sources] ·

T-Bank has released T-Search, an open-source agent-retriever designed for complex, multi-step search tasks. Built on the Qwen3.6-35B-A3B model and trained entirely on synthetic data, T-Search outputs a ranked list of chunks rather than a final answer, allowing integration with various generative models. The release also includes two new benchmarks: TRuST for Russian-language agent search and SynthComp for synthetic search tasks in Russian and English. AI

IMPACT Provides a specialized agent for complex search, potentially improving information retrieval in various applications.

RANK_REASON Release of a new open-source model and associated benchmarks. [lever_c_demoted from research: ic=1 ai=1.0]

Read on Mastodon — fosstodon.org →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

T-Bank releases T-Search agent for multi-step search

How we ranked this

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
Release of a new open-source model and associated benchmarks. [lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
model release, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
58 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

Full methodology in our editorial standards.

COVERAGE [1]

  1. Mastodon — fosstodon.org TIER_1 Русский(RU) · [email protected] ·

    T-Search: How We Trained an Agent for Multi-Step Search Hello everyone! Tolya Potapov here, MLE at T-Bank. We continue to develop the Gen-T line of open models. We

    T-Search: как мы обучали агента многошаговому поиску Всем привет! На связи Толя Потапов, MLE в Т-Банке. Продолжаем развивать линейку открытых моделей Gen-T. Мы делились русскоязычными семействами моделей T-Lite и T-Pro. Сегодня выпускаем модель другого класса — T-Search. Это откр…

    LINKS habr.com/…/1060262

RELATED ENTITIES

RELATED TOPICS