SShogo Bench

Qwen3 30B A3B Instruct 2507

Qwen instruction model for multilingual chat, reasoning, and tool use

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
17
estimated
Long-horizon autonomy
28
estimated
Knowledge-work agents
16
estimated
Computer use
6
estimated
Security & context learning
16
estimated

Price

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

Cheapest host for agents: DekaLLM at $0.10/M.

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

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verifiedest.8%predicted 50
Terminal-Bench 2.0est.8%predicted 8
DeepSWEest.6%predicted 9
FrontierSWEest.5%predicted 1
CursorBenchest.4%predicted 17
GSO-Benchest.4%predicted 2
METR Time Horizonsest.10%predicted 50
Vending-Bench 2est.6%predicted 2
PostTrainBenchest.4%predicted 11
APEX-Agentsest.8%predicted 11
GDPvalest.7%predicted 17
Remote Labor Indexest.5%predicted 0
DeepResearch Benchest.5%predicted 41
OSWorld 2.0est.6%predicted 1
OSWorldest.4%predicted 15
Cybenchest.5%predicted 20
CL-benchest.5%predicted 12

Other benchmarks

BenchmarkCategoryScoreSource
DTBenchKnowledge67.2%source
GPQA DiamondReasoning55.6%source
LMCAKnowledge22.4%source
OTIS Mock AIMEMath62.2%source

API hosts

Live prices and uptime for every host serving Qwen3 30B A3B Instruct 2507 via OpenRouter.

HostInputOutputCache readAgent $/MUptime 24hQuant
DekaLLM$0.090$0.30—$0.1099.0%—
SiliconFlow$0.090$0.30—$0.1099.2%fp8
Nebius$0.10$0.30$0.10$0.1196.7%fp8