SShogo Bench

Claude 3.5 Haiku

Fast Claude model for responsive assistance, classification, and lightweight agents

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
10
estimated
Long-horizon autonomy
20
estimated
Knowledge-work agents
11
estimated
Computer use
2
estimated
Security & context learning
8
estimated

Price

Input
$0.68/M
Output
$3.40/M
Cache read
$0.068/M
3:1 blend
$1.36/M
Agent-effective
$0.31/M

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

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verifiedest.8%predicted 34
Terminal-Bench 2.0est.8%predicted 2
DeepSWEest.6%predicted 3
FrontierSWEest.5%predicted 0
CursorBenchest.4%predicted 11
GSO-Benchest.4%predicted 0
METR Time Horizonsest.10%predicted 36
Vending-Bench 2est.6%predicted 0
PostTrainBenchest.4%predicted 7
APEX-Agentsest.8%predicted 4
GDPvalest.7%predicted 9
Remote Labor Indexest.5%predicted 0
DeepResearch Benchest.5%predicted 38
OSWorld 2.0est.6%predicted 0
OSWorldest.4%predicted 4
Cybenchest.5%predicted 7
CL-benchest.5%predicted 9

Other benchmarks

BenchmarkCategoryScoreSource
Aider PolyglotCoding28.0%source
BALROGLong-horizon19.3%source
CritPtReasoning0.0%source
DTBenchKnowledge56.7%source
GeoBenchMultimodal34.0%source
GPQA DiamondReasoning38.1%source
MATH Level 5Math46.4%source
MMLULegacy74.3%source
OTIS Mock AIMEMath4.3%source
SciCodeCoding27.4%source