PulseAugur
EN
LIVE 10:04:02

New 2.6B Tajik model excels at tool calling but degrades English skills

By PulseAugur Editorial · [1 sources] ·

Researchers have developed Saidzoda-Tajik-LFM-2.6B, a 2.6 billion parameter hybrid model adapted for the Tajik language, focusing on tool-calling capabilities. The adaptation process involved continued pretraining on Tajik web documents followed by supervised fine-tuning on tool-call data. While the model significantly improved Tajik tool-calling accuracy, it notably degraded the model's ability to perform tool calls in its native English format and also caused a decline in English MMLU scores. The researchers caution that the model is not a general Tajik assistant and its calls require verification before execution. AI

IMPACT This research demonstrates a method for adapting existing models to new languages and tasks, highlighting trade-offs in performance across different capabilities.

RANK_REASON The cluster describes a new research paper detailing the adaptation of a language model for a specific language and task, including methodology and evaluation results. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

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

New 2.6B Tajik model excels at tool calling but degrades English skills

How we ranked this

Signal score
1 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
The cluster describes a new research paper detailing the adaptation of a language model for a specific language and task, including methodology and evaluation results. [lever_c_demoted from researc…
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, paper, 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
1 days old
Coverage has settled into its steady-state source set.

Full methodology in our editorial standards.

COVERAGE [1]

  1. arXiv cs.CL TIER_1 English(EN) · Tohir Saidzoda, Nekruz Maxudov, Bedil Karimov, Abubakr Abdullozoda, Amin Jalilov, Jonibek Kishvarov, Rahmonali Sharipov, Mariyamkhon Shodyeva ·

    Saidzoda-Tajik-LFM-2.6B: Adapting a 2.6B Hybrid Model to Tajik Tool Calling, and What Continued Pretraining Takes Away

    arXiv:2610.41576v1 Announce Type: new Abstract: We release Saidzoda-Tajik-LFM-2.6B, a Tajik adaptation of the 2.6B-parameter hybrid model LFM2.5-2.6B-Base, built for tool calling in Tajik on modest hardware. The recipe has two stages: LoRA continued pretraining (CPT) on 1,485,319…

RELATED ENTITIES

RELATED TOPICS