hebrew-llm-eval-suite

A test suite for comparing language models on Hebrew tasks, including reading questions, sentiment, translation, summarization, and Israeli cultural knowledge. It uses published Hebrew datasets and benchmarks.

In plain words
What is it for?
Use it to compare API-based chat models for search, document question-answering, support, research, translation, or summarization in Hebrew. Choose tests that match the product's needs.
Why use it?
It replaces informal guesses and marketing claims with comparable tests for Hebrew products. This helps teams spot models that work poorly for native Hebrew users before launch.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/squadcodercom/squadcoder/hebrew-llm-eval-suite
Any agent
npx skills add squadcodercom/squadcoder --skill hebrew-llm-eval-suite
Clone the repo
git clone --depth 1 https://github.com/squadcodercom/squadcoder

Made for: Claude Code, Codex.

Per session 261 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 5,432 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00261 $0.05432
Opus 5 $0.00130 $0.02716
Sonnet 5 $0.00052 $0.01086
Haiku 4.5 $0.00026 $0.00543

Measured 2d ago against content hash 30dd1f4ba823, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

hebrew-llm-eval-suite scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

The scan reads SKILL.md. This mod also ships 3 executable files (scripts/make_scorecard.py, scripts/run_eval.py, scripts/score_results.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.squadcoder/skills/hebrew-llm-eval-suite/SKILL.md · 257 lines

How it starts

The opening of the file, as written. The whole thing — 257 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Hebrew LLM Eval Suite

Problem

Israeli product teams pick LLMs blind. There is no standardized Hebrew benchmark that a PM can run in an afternoon to compare Claude against GPT against DictaLM against AI21 Jamba on their actual use case. The HuggingFace Open Hebrew LLM Leaderboard exists but is built for base models and few-shot prompts, not for API-hosted chat models. DictaLM publishes benchmark results but only for its own suite. Teams end up guessing, testing informally, or trusting marketing claims. The result is costly model switches after launch, or shipping Hebrew products on models that silently fail on native speakers.

Instructions

Step 1: Pick the right benchmark set for your task

Different benchmarks test different things. Choose the smallest set that covers your actual use case.

Benchmark HuggingFace ID What it tests When to use
HeQ (Hebrew Question Answering) Etelis/HeQ_v1 (HF mirror); canonical at github.com/NNLP-IL/Hebrew-Question-Answering-Dataset Reading comprehension, extractive QA on Hebrew Wikipedia and Geektime articles. 30,147 questions Any product that answers questions over Hebrew text: search, RAG, support, research assistants
HebrewSentiment HebArabNlpProject/HebrewSentiment Sentiment classification (positive, negative, neutral). 41,305 samples. License CC-BY-4.0 Social media analysis, review classification, product feedback
Hebrew Winograd Community port of Winograd Schema Challenge (cs.ubc.ca/~vshwartz/resources/winograd_he.jsonl) Pronoun resolution requiring world knowledge. Reasoning-heavy Any product that needs nuanced Hebrew understanding
NeuLabs-TedTalks (translation) OPUS NeuLab-TedTalks en-he subset English to Hebrew and Hebrew to English translation quality Translation products, multilingual apps
HebNLI HebArabNlpProject/HebNLI Natural Language Inference in Hebrew Classification, content moderation, logical reasoning
HEBREW-MMLU (general knowledge) Hebrew-translated MMLU subset, used by the Open Hebrew LLM Leaderboard ecosystem (verify the active HF mirror before use; openai/MMMLU covers 14 languages but Hebrew is not in the official set) 14-subject general-knowledge accuracy; Hebrew adaptation of Massive Multitask Language Understanding General-purpose chat/RAG products that need broad world knowledge in Hebrew
DictaLM 3.0 Summarization Dicta benchmark suite (see DictaLM 3.0 technical report) Abstractive summarization of Hebrew news Summarization tools, executive briefings
DictaLM 3.0 Nikud Dicta benchmark suite Adding vowel diacritics to unvocalized Hebrew Educational tools, TTS preprocessing, religious text tools
DictaLM 3.0 Israeli Trivia Dicta benchmark suite Knowledge of Israeli culture, geography, history, politics Consumer products where cultural grounding matters

Read the full file on GitHub · 257 lines

Files

What ships with it

10 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 257 lines · 261 tokens per session scan A 30dd1f4ba823

Subscribe to this mod's changes

hebrew-llm-eval-suite is a skill published in the GitHub repository squadcodercom/squadcoder (11 stars, last pushed 2mo ago), licensed MIT. It adds 261 tokens to every session and 5,432 once invoked, about $0.0013 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

product-capability

Translate PRD intent, roadmap asks, or product discussions into an implementation-ready capability plan that exposes constraints, invariants, interfaces, and unresolved decisions before multi-service work starts. Use when the user needs an ecc-native PRD-to-SRS lane instead of vague planning prose.

mturac/everything-openai-codex · 59 tokens

planning-with-files-ar

تخطيط مستمر قائم على الملفات لعمل وكلاء الذكاء الاصطناعي متعدد الخطوات. يحتفظ بملفات taskplan.md و findings.md و progress.md على القرص، وتحقن خطافات دورة الحياة سياق التخطيط المحدد للمشروع. تقرأ الاستعادة التلقائية ملفات تخطيط المشروع فقط. يمكن للأمر الصريح session-catchup.py --metadata فحص بيانات وصفية لجلسات الوكيل…

OthmanAdi/planning-with-files · 189 tokens

kl-consistency-test

Write, calibrate, and debug the prefill-vs-decode logprob (KL) consistency tests in sglang -- the two independent conditions a zero requires (every operator batch-invariant, and the two paths computing the same function), which helper separates them, how to pick a threshold once they hold, and how to localize a…

sgl-project/sglang · 108 tokens

i18n-localization

Internationalization and localization patterns. Detecting hardcoded strings, managing translations, locale files, RTL support.

vudovn/ag-kit · 27 tokens

baoyu-youtube-transcript

Downloads YouTube video transcripts/subtitles and cover images by URL or video ID. Supports multiple languages, translation, chapters, and speaker identification. Caches raw data for fast re-formatting. Use when user asks to "get YouTube transcript", "download subtitles", "get captions", "YouTube字幕", "YouTube封面"…

JimLiu/baoyu-skills · 107 tokens

indication-dossier

Build a source-backed biomedical indication dossier. Use when a research task asks for disease biology, target rationale, patient segmentation, biomarkers, trials, drugs, competitive landscape, or translational evidence.

companion-inc/feynman · 43 tokens