SShogo Bench

Grok 4.6

xAI's frontier model for long-running agents, coding, knowledge work, and visual projects

Shogo Agent Index

Supported#22 of 92
56.2range 48.3–64 · 29% of weight measured
Agentic coding
61
43% measured
Long-horizon autonomy
68
30% measured
Knowledge-work agents
47
32% measured
Computer use
42
estimated
Security & context learning
52
estimated

Price

Input
$2.00/M
Output
$6.00/M
Cache read
$0.50/M
3:1 blend
$3.00/M
Agent-effective
$0.98/M

Cheapest host for agents: xAI at $0.98/M.

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

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verifiedest.8%predicted 81
Terminal-Bench 2.0est.8%predicted 81
DeepSWE67.5%6%mini-swe-agent · $3.45/task
FrontierSWE25.3%5%proximus · $243.43/task
CursorBench41.4%4%source · $6.10/task
GSO-Benchest.4%predicted 39
METR Time Horizonsest.10%predicted 78
Vending-Bench 2$9,0476%source
PostTrainBenchest.4%predicted 28
APEX-Agents65.3%8%source
GDPvalest.7%predicted 53
Remote Labor Indexest.5%predicted 6
DeepResearch Benchest.5%predicted 49
OSWorld 2.0est.6%predicted 14
OSWorldest.4%predicted 83
Cybenchest.5%predicted 81
CL-benchest.5%predicted 23

Other benchmarks

BenchmarkCategoryScoreSource
ARC-AGIReasoning87.5%source
ARC-AGI-2Reasoning67.1%source
Chess PuzzlesReasoning33.0%source
CritPtReasoning19.7%source
DTBenchKnowledge97.3%source
FrontierCodeCoding48.0%grok-build
FrontierMath Tiers 1-3Math66.0%source
FrontierMath Tier 4Math31.7%source
GPQA DiamondReasoning94.0%source
LMCAKnowledge48.5%source
MirrorCodeCoding41.1%source
OTIS Mock AIMEMath99.2%source
ProofBenchMath51.0%source
SciCodeCoding56.5%source
SimpleBenchReasoning75.9%source
SimpleQA VerifiedKnowledge49.3%source
WebDev ArenaCoding1618source

API hosts

Live prices and uptime for every host serving Grok 4.6 via OpenRouter.

HostInputOutputCache readAgent $/MUptime 24hQuant
xAI$2.00$6.00$0.50$0.9899.6%—
xAI$2.00$6.00$0.50$0.9899.7%—
xAI$4.00$12.00$1.00$1.95——
xAI$4.00$12.00$1.00$1.95——