SShogo Bench

MiMo-V2.6-Pro

MiMo Pro model for multimodal coding agents and long-context automation

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
65
estimated
Long-horizon autonomy
69
estimated
Knowledge-work agents
45
32% measured
Computer use
47
estimated
Security & context learning
55
estimated

Price

Input
$0.43/M
Output
$0.87/M
Cache read
$0.004/M
3:1 blend
$0.54/M
Agent-effective
$0.11/M

Cheapest host for agents: GMICloud at $0.080/M.

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

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verifiedest.8%predicted 83
Terminal-Bench 2.0est.8%predicted 87
DeepSWEest.6%predicted 64
FrontierSWEest.5%predicted 32
CursorBenchest.4%predicted 45
GSO-Benchest.4%predicted 50
METR Time Horizonsest.10%predicted 81
Vending-Bench 2est.6%predicted 73
PostTrainBenchest.4%predicted 31
APEX-Agents54.4%8%source
GDPvalest.7%predicted 58
Remote Labor Indexest.5%predicted 8
DeepResearch Benchest.5%predicted 50
OSWorld 2.0est.6%predicted 20
OSWorldest.4%predicted 88
Cybenchest.5%predicted 86
CL-benchest.5%predicted 25

Other benchmarks

BenchmarkCategoryScoreSource
CritPtReasoning26.6%source
ProofBenchMath70.0%source
SciCodeCoding60.9%source
WebDev ArenaCoding1618source

API hosts

Live prices and uptime for every host serving MiMo-V2.6-Pro via OpenRouter.

HostInputOutputCache readAgent $/MUptime 24hQuant
GMICloud$0.33$0.65$0.003$0.08094.7%bf16
DeepInfra$0.39$0.78$0.003$0.10100.0%fp8
StreamLake$0.43$0.87$0.004$0.1198.9%—
Xiaomi$0.43$0.87$0.004$0.1199.9%fp8
PrimaLabs$2.64$5.28$0.024$0.6599.7%fp8