SShogo Bench

Llama 4 Scout

Llama 4 Scout 17B Instruct (16E) is a mixture-of-experts (MoE) language model developed by Meta, activating 17 billion parameters out of a total of 109B. It supports native multimodal input…

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
12
estimated
Long-horizon autonomy
22
estimated
Knowledge-work agents
13
estimated
Computer use
3
estimated
Security & context learning
10
estimated

Price

Input
$0.10/M
Output
$0.30/M
Cache read
—/M
3:1 blend
$0.15/M
Agent-effective
$0.11/M

Cheapest host for agents: DeepInfra at $0.090/M.

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

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verifiedest.8%predicted 39
Terminal-Bench 2.0est.8%predicted 3
DeepSWEest.6%predicted 4
FrontierSWEest.5%predicted 0
CursorBenchest.4%predicted 12
GSO-Benchest.4%predicted 1
METR Time Horizonsest.10%predicted 40
Vending-Bench 2est.6%predicted 1
PostTrainBenchest.4%predicted 8
APEX-Agentsest.8%predicted 6
GDPvalest.7%predicted 10
Remote Labor Indexest.5%predicted 0
DeepResearch Benchest.5%predicted 39
OSWorld 2.0est.6%predicted 0
OSWorldest.4%predicted 6
Cybenchest.5%predicted 9
CL-benchest.5%predicted 10

Other benchmarks

BenchmarkCategoryScoreSource
ARC-AGIReasoning0.5%source
ARC-AGI-2Reasoning0.0%source
CritPtReasoning0.0%source
DTBenchKnowledge57.9%source
Fiction.LiveBenchLong context36.0%source
GPQA DiamondReasoning51.8%source
LMCAKnowledge12.0%source
MATH Level 5Math62.3%source
OTIS Mock AIMEMath7.8%source
SciCodeCoding21.3%source

API hosts

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

HostInputOutputCache readAgent $/MUptime 24hQuant
DeepInfra$0.080$0.24—$0.09095.4%fp8
DeepInfra$0.10$0.30—$0.11100.0%fp8
DeepInfra$0.15$0.45—$0.1699.1%fp8