Compare top models

LLM Benchmarks

See how today's leading models write Spacetime code, scored task by task on live modules.

Anthropic: Claude Fable 5.1

TypeScript · With Guidelines · 2026-09-01
100.0%Task Pass Rate
89/89Checks Passed
47Total Evals

auth

100% · 8 evals
PassTaskChecksRun Time
Auth Identity Check1/17.4s
Private Vs Public Table1/17.3s
Registered User Gate1/110.3s
Admin Bootstrap1/19.0s
Role Based Access1/18.8s
Ban List1/110.1s
Rate Limit1/19.2s
Shared Document1/110.5s

basics

100% · 15 evals
PassTaskChecksRun Time
Empty Reducers1/18.4s
Basic Tables1/17.0s
Scheduled Table1/17.8s
Struct In Table1/15.9s
Insert2/26.8s
Update3/310.1s
Delete3/37.0s
Crud4/48.3s
Index Lookup3/310.5s
Init4/47.3s
Connect1/17.6s
Helper Function3/36.9s
Schedule At Time1/18.8s
Cancel Schedule1/19.2s
Lifecycle Player1/17.3s

data modeling

100% · 6 evals
PassTaskChecksRun Time
Event Table1/18.4s
Optional Fields1/18.5s
Cascade Delete1/18.6s
Filter And Aggregate1/110.7s
Wo Table Join1/110.7s
Unique Constraint1/18.4s

queries

100% · 8 evals
PassTaskChecksRun Time
View Basic1/18.1s
View Per User1/17.5s
Range Query1/18.5s
Sort And Limit1/18.8s
Find First1/18.4s
Select Distinct1/18.7s
Count Without Collect1/111.0s
Multi Column Filter1/19.7s

schema

100% · 10 evals
PassTaskChecksRun Time
Spacetime Product Type3/39.5s
Spacetime Sum Type3/38.8s
Elementary Columns3/38.8s
Product Type Columns3/38.9s
Sum Type Columns3/310.7s
Scheduled Columns2/210.5s
Constraints3/311.3s
Many To Many5/511.4s
Ecs7/713.9s
Multi Column Index5/510.8s