SShogo Bench

Qwen3 235B A22B 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
21
estimated
Long-horizon autonomy
31
estimated
Knowledge-work agents
19
estimated
Computer use
9
estimated
Security & context learning
20
estimated

Price

Input
$0.20/M
Output
$0.60/M
Cache read
$0.20/M
3:1 blend
$0.30/M
Agent-effective
$0.22/M

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

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verifiedest.8%predicted 55
Terminal-Bench 2.0est.8%predicted 13
DeepSWEest.6%predicted 13
FrontierSWEest.5%predicted 1
CursorBenchest.4%predicted 20
GSO-Benchest.4%predicted 3
METR Time Horizonsest.10%predicted 54
Vending-Bench 2est.6%predicted 4
PostTrainBenchest.4%predicted 13
APEX-Agentsest.8%predicted 14
GDPvalest.7%predicted 20
Remote Labor Indexest.5%predicted 0
DeepResearch Benchest.5%predicted 42
OSWorld 2.0est.6%predicted 1
OSWorldest.4%predicted 22
Cybenchest.5%predicted 27
CL-benchest.5%predicted 14

Other benchmarks

BenchmarkCategoryScoreSource
Aider PolyglotCoding59.6%source
ARC-AGIReasoning11.0%source
ARC-AGI-2Reasoning1.3%source
DTBenchKnowledge78.4%source
Fiction.LiveBenchLong context52.9%source
LMCAKnowledge23.5%source