SShogo Bench

Qwen3.8 Max Preview

Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows

Shogo Agent Index

Estimated#30 of 92
53.6range 41.7–65.5 · 19% of weight measured
Agentic coding
58
31% measured
Long-horizon autonomy
63
estimated
Knowledge-work agents
46
32% measured
Computer use
41
estimated
Security & context learning
52
estimated

Price

Input
Free/M
Output
Free/M
Cache read
Free/M
3:1 blend
Free/M
Agent-effective
Free/M

Cheapest host for agents: Alibaba at $1.26/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 80
DeepSWE57.5%6%mini-swe-agent · $3.73/task
FrontierSWE15.8%5%proximus · $55.14/task
CursorBenchest.4%predicted 41
GSO-Benchest.4%predicted 38
METR Time Horizonsest.10%predicted 78
Vending-Bench 2est.6%predicted 61
PostTrainBenchest.4%predicted 28
APEX-Agents63.3%8%source
GDPvalest.7%predicted 52
Remote Labor Indexest.5%predicted 6
DeepResearch Benchest.5%predicted 49
OSWorld 2.0est.6%predicted 14
OSWorldest.4%predicted 82
Cybenchest.5%predicted 80
CL-benchest.5%predicted 23

Other benchmarks

BenchmarkCategoryScoreSource
Chess PuzzlesReasoning46.0%source
CritPtReasoning20.0%source
DTBenchKnowledge92.0%source
FrontierMath Tiers 1-3Math74.7%source
FrontierMath Tier 4Math46.3%source
GPQA DiamondReasoning92.7%source
LMCAKnowledge46.2%source
MirrorCodeCoding52.2%source
OTIS Mock AIMEMath99.4%source
ProofBenchMath58.0%source
SciCodeCoding53.2%source
SimpleQA VerifiedKnowledge45.8%source
WebDev ArenaCoding1671source

API hosts

Live prices and uptime for every host serving Qwen3.8 Max Preview via OpenRouter.

HostInputOutputCache readAgent $/MUptime 24hQuant
Alibaba$1.15$3.45—$1.26——