SShogo Bench

Phi 4

[Microsoft Research](/microsoft) Phi-4 is designed to perform well in complex reasoning tasks and can operate efficiently in situations with limited memory or where quick responses are needed. At 14 billion…

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
13
estimated
Long-horizon autonomy
23
estimated
Knowledge-work agents
13
estimated
Computer use
3
estimated
Security & context learning
11
estimated

Price

Input
$0.070/M
Output
$0.14/M
Cache read
—/M
3:1 blend
$0.090/M
Agent-effective
$0.070/M

Cheapest host for agents: DeepInfra at $0.060/M.

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

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verifiedest.8%predicted 41
Terminal-Bench 2.0est.8%predicted 3
DeepSWEest.6%predicted 5
FrontierSWEest.5%predicted 0
CursorBenchest.4%predicted 13
GSO-Benchest.4%predicted 1
METR Time Horizonsest.10%predicted 42
Vending-Bench 2est.6%predicted 1
PostTrainBenchest.4%predicted 9
APEX-Agentsest.8%predicted 6
GDPvalest.7%predicted 12
Remote Labor Indexest.5%predicted 0
DeepResearch Benchest.5%predicted 39
OSWorld 2.0est.6%predicted 0
OSWorldest.4%predicted 8
Cybenchest.5%predicted 11
CL-benchest.5%predicted 11

Other benchmarks

BenchmarkCategoryScoreSource
BALROGLong-horizon11.6%source
CritPtReasoning0.0%source
GPQA DiamondReasoning56.1%source
MATH Level 5Math64.9%source
MMLULegacy84.8%source
OTIS Mock AIMEMath13.8%source

API hosts

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

HostInputOutputCache readAgent $/MUptime 24hQuant
DeepInfra$0.056$0.11—$0.060100.0%bf16
DeepInfra$0.070$0.14—$0.070100.0%bf16