Atlas Graph Explorer
Wiki
Graph
Edges
Home
Benchmark
benchmark:gsm8k
GSM8K
benchmark:gsm8k
Benchmark
benchmarks/benchmarks/benchmarks-math.yaml
·
Open in Graph →
overview
json
graph
Attributes
displayName
GSM8K
homepageUrl
https://github.com/openai/grade-school-math
kind
math
targetsKind
ModelVersion
description
GSM8K (OpenAI) is 8.5K linguistically diverse grade-school math word problems requiring multi-step arithmetic reasoning.
Outgoing edges
(1)
covers
1
skill-area:mathematical-reasoning
·
SkillArea
Mathematical Reasoning
Incoming edges
(4)
belongs_to_benchmark
1
test-set:gsm8k-test
·
TestSet
GSM8K test split
bounds_subject
1
scope-boundary:gsm8k.scope
·
ScopeBoundary
for_benchmark
2
eval-run:gsm8k.gemma-2-27b.2024-06
·
EvalRun
eval-run:gsm8k.claude-sonnet-4-5.2025-09
·
EvalRun