Computation and Language · arXiv:2510.25409 v2 · submitted 29 Oct 2025
BhashaBench V1: A Comprehensive Benchmark for the Quadrant of Indic Domains
The paper says
“Evaluation of 29+ LLMs reveals significant domain and language specific performance gaps, with especially large disparities in low-resource domains.”
from the abstract · picked out as the main claim by a classifier
Reviews
Reviews by people
No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.
The paper
Abstract
The rapid advancement of large language models(LLMs) has intensified the need for domain and culture specific evaluation. Existing benchmarks are largely Anglocentric and domain-agnostic, limiting their applicability to India-centric contexts. To address this gap, we introduce BhashaBench V1, the first domain-specific, multi-task, bilingual benchmark focusing on critical Indic knowledge systems. BhashaBench V1 contains 74,166 meticulously curated question-answer pairs, with 52,494 in English and 21,672 in Hindi, sourced from authentic government and domain-specific exams. It spans four major domains: Agriculture, Legal, Finance, and Ayurveda, comprising 90+ subdomains and covering 500+ topics, enabling fine-grained evaluation. Evaluation of 29+ LLMs reveals significant domain and language specific performance gaps, with especially large disparities in low-resource domains. For instance, GPT-4o achieves 76.49% overall accuracy in Legal but only 59.74% in Ayurveda. Models consistently perform better on English content compared to Hindi across all domains. Subdomain-level analysis shows that areas such as Cyber Law, International Finance perform relatively well, while Panchakarma, Seed Science, and Human Rights remain notably weak. BhashaBench V1 provides a comprehensive dataset for evaluating large language models across India's diverse knowledge domains. It enables assessment of models' ability to integrate domain-specific knowledge with bilingual understanding. All code, benchmarks, and resources are publicly available to support open research.
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. No sentence in this record was written by a model.
abstract compiled · 2026-11-16 · no language-model review
How the main claim is stated: asserted 1.00
What the abstract itself states, as probabilities: claims first 0.94 · states limitation 0.23 · human participants 0.11 · controlled comparison 0.09 · randomized assignment 0.01 · states quantitative result 0.99 · compares with prior results 0.41 · mentions code or data availability 0.99
The abstract, sentence by sentence
This table labels the abstract only (arXiv listing abstract v2). The number is the probability the classifier gave the label.
| Sentence | Role | The authors' own work | States the main conclusion | Strength |
|---|---|---|---|---|
| The rapid advancement of large language models(LLMs) has intensified the need for domain and culture specific evaluation. | background 1.00 | 0.05 | 0.03 | asserted 1.00 |
| Existing benchmarks are largely Anglocentric and domain-agnostic, limiting their applicability to India-centric contexts. | gap 0.90 | 0.05 | 0.03 | asserted 1.00 |
| To address this gap, we introduce BhashaBench V1, the first domain-specific, multi-task, bilingual benchmark focusing on critical Indic knowledge systems. | contribution 1.00 | 0.92 | 0.08 | asserted 1.00 |
| BhashaBench V1 contains 74,166 meticulously curated question-answer pairs, with 52,494 in English and 21,672 in Hindi, sourced from authentic government and domain-specific exams. | method 0.76 | 0.93 | 0.04 | asserted 1.00 |
| It spans four major domains: Agriculture, Legal, Finance, and Ayurveda, comprising 90+ subdomains and covering 500+ topics, enabling fine-grained evaluation. | contribution 0.53 | 0.85 | 0.04 | asserted 1.00 |
| Evaluation of 29+ LLMs reveals significant domain and language specific performance gaps, with especially large disparities in low-resource domains. | result 0.99 | 0.94 | 0.57 | asserted 1.00 |
| For instance, GPT-4o achieves 76.49% overall accuracy in Legal but only 59.74% in Ayurveda. | result 1.00 | 0.92 | 0.09 | asserted 1.00 |
| Models consistently perform better on English content compared to Hindi across all domains. | result 1.00 | 0.90 | 0.15 | asserted 1.00 |
| Subdomain-level analysis shows that areas such as Cyber Law, International Finance perform relatively well, while Panchakarma, Seed Science, and Human Rights remain notably weak. | result 0.99 | 0.93 | 0.11 | asserted 0.99 |
| BhashaBench V1 provides a comprehensive dataset for evaluating large language models across India's diverse knowledge domains. | contribution 0.94 | 0.71 | 0.23 | asserted 1.00 |
| It enables assessment of models' ability to integrate domain-specific knowledge with bilingual understanding. | contribution 0.82 | 0.33 | 0.11 | asserted 0.99 |
| All code, benchmarks, and resources are publicly available to support open research. | contribution 0.85 | 0.65 | 0.03 | asserted 1.00 |
What kind of paper the classifier reports this is: dataset or benchmark 1.00 · evidence from observation or experiment 0.91
What was examined, and what was not
Sources read
- arXiv listing abstract v2
Not examined
- full text
- figures and tables
- references and citing sentences
- relations to other papers
Levels run
- P abstract
Not run
- S sections
- T body sentences
- C citations
- R cross-paper relations
How the labels were made
Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-11-16: 1 request, 8,835 input tokens.
Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. This record as JSON
How the literature uses it
3 arXiv papers cite this in 4 sentences
Papers citing it · 1 sentence without a confident label
Uses its data or values 1 paper
-
The 11 datasets cover four task categories: * Classification: Apt-eval [ref] (safety/robustness), GT-FintechLab [ref] (finance), and BhashaBench-Finance [ref] (multilingual finance). * Retrieval: AIR-Bench-Finance [ref] , LIMIT [ref] (instruction-following), and ArXiv-Abstracts 2025 [<https://huggingface.co/datasets/almanach/arxiv_abstracts_2025>] (scientific literature).
FLARE: Task-agnostic embedding model evaluation through a normalization process · Embedders and Datasets · 0.90
Cites it as background 1 paper
-
Several benchmarks assess LLM capabilities in this setting: IndicMMLU-Pro [ref] extends MMLU-Pro to nine Indic languages for comprehension, reasoning, and generation, while BhashaBench-V1 [ref] evaluates India-centric knowledge in English and Hindi across domains such as Agriculture, Legal, Finance, and Ayurveda.
SurakshaEval: An Indic Safety Benchmark for Multilingual LLMs · Related Work · 0.89
Challenges and discussion
Know something about this paper? Leave a comment.
Come back by email
A short list of the week's newly reviewed papers. Nothing is sent until you confirm from your inbox, and every email has a link that stops it. No account.
No comments yet.
Sign in to comment