SShogo Bench

GPT-4

GPT model for general reasoning, writing, coding, and tool-assisted tasks

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
8
estimated
Long-horizon autonomy
19
50% measured
Knowledge-work agents
10
estimated
Computer use
1
estimated
Security & context learning
6
estimated

Price

Input
$30.00/M
Output
$60.00/M
Cache read
—/M
3:1 blend
$37.50/M
Agent-effective
$31.43/M

Cheapest host for agents: Azure at $31.43/M.

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

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verifiedest.8%predicted 28
Terminal-Bench 2.0est.8%predicted 1
DeepSWEest.6%predicted 2
FrontierSWEest.5%predicted 0
CursorBenchest.4%predicted 8
GSO-Benchest.4%predicted 0
METR Time Horizons36.1%10%source
Vending-Bench 2est.6%predicted 0
PostTrainBenchest.4%predicted 6
APEX-Agentsest.8%predicted 3
GDPvalest.7%predicted 6
Remote Labor Indexest.5%predicted 0
DeepResearch Benchest.5%predicted 36
OSWorld 2.0est.6%predicted 0
OSWorldest.4%predicted 2
Cybenchest.5%predicted 4
CL-benchest.5%predicted 8

Other benchmarks

BenchmarkCategoryScoreSource
DTBenchKnowledge62.7%source
GPQA DiamondReasoning35.7%source
LMCAKnowledge17.1%source
MATH Level 5Math23.0%source
MMLULegacy86.4%source
OTIS Mock AIMEMath1.1%source

API hosts

Live prices and uptime for every host serving GPT-4 via OpenRouter.

HostInputOutputCache readAgent $/MUptime 24hQuant
Azure$30.00$60.00—$31.43100.0%—