SShogo Bench

Qwen3.6 27B

Qwen vision-language model for visual reasoning, documents, and agent tasks

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
35
estimated
Long-horizon autonomy
42
estimated
Knowledge-work agents
27
estimated
Computer use
22
estimated
Security & context learning
35
estimated

Price

Input
$0.60/M
Output
$3.00/M
Cache read
$0.30/M
3:1 blend
$1.20/M
Agent-effective
$0.47/M

Cheapest host for agents: Chutes at $0.15/M.

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

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verifiedest.8%predicted 68
Terminal-Bench 2.0est.8%predicted 40
DeepSWEest.6%predicted 28
FrontierSWEest.5%predicted 5
CursorBenchest.4%predicted 28
GSO-Benchest.4%predicted 10
METR Time Horizonsest.10%predicted 66
Vending-Bench 2est.6%predicted 17
PostTrainBenchest.4%predicted 18
APEX-Agentsest.8%predicted 27
GDPvalest.7%predicted 33
Remote Labor Indexest.5%predicted 1
DeepResearch Benchest.5%predicted 45
OSWorld 2.0est.6%predicted 3
OSWorldest.4%predicted 50
Cybenchest.5%predicted 52
CL-benchest.5%predicted 18

Other benchmarks

BenchmarkCategoryScoreSource
CritPtReasoning0.9%source
DTBenchKnowledge78.1%source
FrontierMath Tiers 1-3Math35.1%source
GPQA DiamondReasoning85.9%source
LMCAKnowledge34.5%source
OTIS Mock AIMEMath91.1%source
SciCodeCoding37.3%source

API hosts

Live prices and uptime for every host serving Qwen3.6 27B via OpenRouter.

HostInputOutputCache readAgent $/MUptime 24hQuant
Chutes$0.25$2.00$0.025$0.1597.8%fp8
DeepInfra$0.26$2.56—$0.37100.0%fp8
SiliconFlow$0.30$3.20—$0.4498.7%fp8
DeepInfra$0.32$3.20—$0.4699.8%fp8
Venice$0.33$3.25—$0.4698.3%fp8
Alibaba$0.45$2.70—$0.5699.8%—
DeepInfra$0.48$4.80—$0.6998.6%fp8
Reka$0.60$6.00$0.50$0.7899.9%—