SShogo Bench

Muse Spark

Muse Spark from Meta.

Shogo Agent Index

Unranked

Not ranked yet: 0 agentic benchmarks measured, and we need at least 2 plus an overall capability score. It will be ranked as results are published.

Agentic coding
49
estimated
Long-horizon autonomy
53
estimated
Knowledge-work agents
36
estimated
Computer use
33
estimated
Security & context learning
46
estimated

Price

No public API price.

Agent-effective = 20 input tokens per output token, 85% cache hits. Why.

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verifiedest.8%predicted 76
Terminal-Bench 2.0est.8%predicted 66
DeepSWEest.6%predicted 44
FrontierSWEest.5%predicted 13
CursorBenchest.4%predicted 35
GSO-Benchest.4%predicted 24
METR Time Horizonsest.10%predicted 73
Vending-Bench 2est.6%predicted 40
PostTrainBenchest.4%predicted 23
APEX-Agentsest.8%predicted 41
GDPvalest.7%predicted 44
Remote Labor Indexest.5%predicted 3
DeepResearch Benchest.5%predicted 47
OSWorld 2.0est.6%predicted 8
OSWorldest.4%predicted 71
Cybenchest.5%predicted 70
CL-benchest.5%predicted 21

Other benchmarks

BenchmarkCategoryScoreSource
CritPtReasoning11.3%source
GPQA DiamondReasoning89.8%source
Humanity's Last ExamReasoning40.6%source
OTIS Mock AIMEMath88.9%source
ProofBenchMath17.0%source
SciCodeCoding51.5%source
WebDev ArenaCoding1513source