Benchmark Almanac

Snowflake

We track 16 benchmarks that Snowflake made or helped make. Its models are on its lab page.

0 in use 6 with public code 7 with a leaderboard here

Benchmarks

Name Kind Released Best score Status
FeatEng Models write Python functions to transform 103 tabular datasets. The score measures how much those changes reduce prediction error for an XGBoost model. Data science and SQL Oct 2024 11.1% Not in cards
Snowflake Text-to-SQL HARD Models generate SQL for difficult questions about unfamiliar enterprise schemas. Snowflake compares models by query accuracy. Data science and SQL May 2026 47.9% Not in cards
Spider 2.0-AIFunc Agents generate SQL that uses AI functions to solve 465 tasks across 125 databases. The evaluator compares execution results with reference results. Data science and SQL Jul 2026 70.3% Not in cards
DARE-Bench Agents write code to produce predictions from structured files. The benchmark checks predictions and adherence to modeling instructions. Data science and SQL Feb 2026 49.3% Not in cards
HyDRA-Bench Agents interleave database queries and document retrieval to answer questions. A model judge checks the final answer. Data science and SQL Aug 2026 53.9% Not in cards
HybridDeepResearch Agents combine database queries and web search to answer research questions under constraints. Data science and SQL Aug 2026 32.8% Not in cards
Arctic-Extract Business Document Benchmark Models answer questions about business documents and extract information from varied layouts. The internal visual tasks include contracts, signatures and checkboxes. Documents, charts and text in images Aug 2025 58.5 Not in cards
AvalancheBench (May study) Agents recover underlying patterns from synthetic enterprise data and report their findings. Data science and SQL May 2026 Not in cards
Cortex Analyst Multi-Turn Question Rewriting Models rewrite a follow-up question so it stands alone. Mistral Large 2 judges whether the rewrite preserves the necessary conversation history. Multi-turn conversation Nov 2024 Not in cards
data-eng-bench Agents build and repair dbt models across 103 retail warehouse tasks. The verifier compares materialized tables row by row against a reference solution. Data science and SQL Aug 2026 Not in cards
Snowflake AI_EXTRACT Document Extraction Evaluation Models extract fields, lists and tables across 2,789 annotated tasks from 11 datasets. Missing items, extra items and incorrect cells reduce the score. Documents, charts and text in images Nov 2026 Not in cards
Snowflake Analytical Search: ArXiv Papers Agents count, group and compare findings across research papers. The evaluation checks numeric and categorical answers against references and scores lists with F1. Deep research Jun 2026 Not in cards
Snowflake Arctic Enterprise Intelligence Models write code, generate SQL and follow instructions. Snowflake averages the coding, SQL and instruction-following scores. Composite indices Apr 2024 Not in cards
Snowflake Financial Filings RAG Evaluation Systems answer about 500 questions over 23,000 SEC filings. Annotators identify supporting passages, pages and answers. Finance Mar 2025 Not in cards
Snowflake Financial PDF Table Parsing Models reconstruct tables from 50 manually annotated financial documents. Evaluators compare complete tables, table dimensions and individual cells with reference Markdown. Documents, charts and text in images Feb 2026 Not in cards
Snowflake Multimodal PDF Retrieval Evaluation Retrievers find relevant pages in technical manuals, SEC filings and presentation slides. The tasks include 143 chart questions and 69 text questions about manuals. Documents, charts and text in images Apr 2025 Not in cards