SShogo Bench

Beam 1.0

Beam 1.0 from Reflection.

Shogo Agent Index

Unranked

Not ranked yet: 1 agentic benchmark measured, and we need at least 2 plus an overall capability score. It will be ranked as results are published.

Agentic coding
43
estimated
Long-horizon autonomy
49
estimated
Knowledge-work agents
37
32% measured
Computer use
29
estimated
Security & context learning
42
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 73
Terminal-Bench 2.0est.8%predicted 57
DeepSWEest.6%predicted 38
FrontierSWEest.5%predicted 9
CursorBenchest.4%predicted 33
GSO-Benchest.4%predicted 18
METR Time Horizonsest.10%predicted 71
Vending-Bench 2est.6%predicted 30
PostTrainBenchest.4%predicted 21
APEX-Agents48.8%8%source
GDPvalest.7%predicted 40
Remote Labor Indexest.5%predicted 2
DeepResearch Benchest.5%predicted 47
OSWorld 2.0est.6%predicted 6
OSWorldest.4%predicted 64
Cybenchest.5%predicted 64
CL-benchest.5%predicted 20

Other benchmarks

BenchmarkCategoryScoreSource
ARC-AGIReasoning74.0%source
ARC-AGI-2Reasoning23.5%source
CritPtReasoning13.1%source
ProofBenchMath18.0%source
SciCodeCoding48.5%source