SShogo Bench

Claude 3.7 Sonnet

Balanced Claude model for coding, analysis, agent workflows, and cost control

Shogo Agent Index

Supported#82 of 92
24.2range 18.6–29.8 · 36% of weight measured
Agentic coding
25
34% measured
Long-horizon autonomy
35
50% measured
Knowledge-work agents
21
20% measured
Computer use
15
40% measured
Security & context learning
18
50% measured

Price

Input
$2.60/M
Output
$13.00/M
Cache read
$0.26/M
3:1 blend
$5.20/M
Agent-effective
$1.20/M

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

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verified61.0%8%source
Terminal-Bench 2.0est.8%predicted 20
DeepSWEest.6%predicted 17
FrontierSWEest.5%predicted 2
CursorBenchest.4%predicted 22
GSO-Bench3.8%4%OpenHands
METR Time Horizons60.0%10%source
Vending-Bench 2est.6%predicted 6
PostTrainBenchest.4%predicted 14
APEX-Agentsest.8%predicted 18
GDPvalest.7%predicted 24
Remote Labor Indexest.5%predicted 1
DeepResearch Bench43.6%5%source
OSWorld 2.0est.6%predicted 1
OSWorld35.8%4%claude-3-7-sonnet-20250219 (50 steps)
Cybench20.0%5%source
CL-benchest.5%predicted 15

Other benchmarks

BenchmarkCategoryScoreSource
Aider PolyglotCoding64.9%source
ARC-AGIReasoning28.6%source
ARC-AGI-2Reasoning0.9%source
Fiction.LiveBenchLong context83.3%source
GeoBenchMultimodal68.0%source
GPQA DiamondReasoning79.7%source
Humanity's Last ExamReasoning8.0%source
MATH Level 5Math91.2%source
OTIS Mock AIMEMath57.8%source
SimpleBenchReasoning46.4%source
TheAgentCompanyKnowledge work30.9%source
VPCTMultimodal39.0%source