SShogo Bench

Gemini 3 Pro Preview

Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts

Shogo Agent Index

Supported#48 of 92
44.7range 37.9–51.6 · 58% of weight measured
Agentic coding
48
57% measured
Long-horizon autonomy
60
80% measured
Knowledge-work agents
34
68% measured
Computer use
33
estimated
Security & context learning
43
50% measured

Price

Input
$2.00/M
Output
$12.00/M
Cache read
—/M
3:1 blend
$4.50/M
Agent-effective
$2.48/M

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

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verified72.9%8%source
Terminal-Bench 2.069.4%8%Ante
DeepSWEest.6%predicted 44
FrontierSWEest.5%predicted 12
CursorBenchest.4%predicted 35
GSO-Bench18.6%4%OpenHands
METR Time Horizons71.0%10%source
Vending-Bench 2$5,4786%source
PostTrainBenchest.4%predicted 23
APEX-Agentsest.8%predicted 41
GDPval40.3%7%source
Remote Labor Index1.3%5%source
DeepResearch Bench46.3%5%source
OSWorld 2.0est.6%predicted 7
OSWorldest.4%predicted 71
Cybenchest.5%predicted 70
CL-bench15.8%5%source

Other benchmarks

BenchmarkCategoryScoreSource
AlgoTuneCoding1.83×source
ARC-AGIReasoning75.0%source
ARC-AGI-2Reasoning31.1%source
BALROGLong-horizon58.1%source
CritPtReasoning6.9%source
GeoBenchMultimodal84.0%source
GPQA DiamondReasoning92.6%source
Humanity's Last ExamReasoning37.5%source
OTIS Mock AIMEMath91.4%source
ProofBenchMath20.0%source
SimpleBenchReasoning76.4%source
VPCTMultimodal91.0%source
WebDev ArenaCoding1439source