SShogo Bench

GLM-5.3

Flagship GLM model for long-horizon coding, agents, and complex project delivery

Shogo Agent Index

Supported#25 of 92
55.2range 47.7–62.7 · 33% of weight measured
Agentic coding
62
43% measured
Long-horizon autonomy
68
50% measured
Knowledge-work agents
43
32% measured
Computer use
40
estimated
Security & context learning
51
estimated

Price

Input
$1.20/M
Output
$4.18/M
Cache read
$0.30/M
3:1 blend
$1.94/M
Agent-effective
$0.61/M

Cheapest host for agents: Wafer at $0.13/M.

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

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verifiedest.8%predicted 80
Terminal-Bench 2.0est.8%predicted 79
DeepSWE69.0%6%mini-swe-agent · $3.99/task
FrontierSWE30.2%5%proximus · $97.22/task
CursorBench42.6%4%source · $5.05/task
GSO-Benchest.4%predicted 36
METR Time Horizonsest.10%predicted 77
Vending-Bench 2$8,1646%source
PostTrainBench34.5%4%Claude Code
APEX-Agents56.6%8%source
GDPvalest.7%predicted 51
Remote Labor Indexest.5%predicted 5
DeepResearch Benchest.5%predicted 49
OSWorld 2.0est.6%predicted 13
OSWorldest.4%predicted 81
Cybenchest.5%predicted 79
CL-benchest.5%predicted 23

Other benchmarks

BenchmarkCategoryScoreSource
Chess PuzzlesReasoning37.0%source
CritPtReasoning19.1%source
DTBenchKnowledge87.7%source
FrontierCodeCoding40.1%devin-cli
FrontierMath Tiers 1-3Math68.8%source
FrontierMath Tier 4Math29.3%source
GPQA DiamondReasoning90.9%source
LMCAKnowledge55.5%source
OTIS Mock AIMEMath91.1%source
ProofBenchMath49.0%source
SciCodeCoding59.0%source
SimpleBenchReasoning66.2%source
SimpleQA VerifiedKnowledge41.0%source
WebDev ArenaCoding1614source

API hosts

Live prices and uptime for every host serving GLM-5.3 via OpenRouter.

HostInputOutputCache readAgent $/MUptime 24hQuant
Wafer$0.070$1.40$0.069$0.13100.0%fp4
Featherless$0.42$1.32$0.078$0.1996.2%—
DeepInfra$0.45$2.00$0.10$0.24100.0%fp4
Wafer$0.18$1.40$0.18$0.24100.0%fp4
Decart$0.56$1.81$0.10$0.25100.0%fp4
InferenceNet$0.084$3.60$0.079$0.25100.0%fp4
InferenceNet$0.084$3.60$0.079$0.25100.0%fp4
Novita$0.63$1.98$0.12$0.28100.0%fp8
Relace$0.070$4.38$0.070$0.28100.0%fp4
Tenstorrent$0.60$2.50$0.10$0.2999.6%fp8
DeepInfra$0.56$2.50$0.13$0.3099.0%fp4
SiliconFlow$0.70$1.90$0.13$0.3099.9%fp8
Cloudflare$0.70$2.20$0.13$0.31100.0%—
Fireworks$0.70$2.20$0.13$0.3194.4%—
GMICloud$0.70$2.20$0.13$0.3199.1%fp8
Cohere$0.50$3.50$0.10$0.3299.9%nvfp4
Sail Research$0.75$3.00$0.12$0.35100.0%fp8
AkashML$0.19$4.40$0.19$0.39100.0%fp4
CoreWeave$0.75$3.00$0.19$0.4099.7%nvfp4
DigitalOcean$0.91$2.86$0.17$0.4099.9%—
Inceptron$1.05$3.40$0.12$0.4199.9%fp4
SambaNova$1.40$4.40$0.070$0.47100.0%—
Sail Research$1.05$4.20$0.17$0.49100.0%fp8
BaseTen$1.40$4.40$0.14$0.52100.0%fp4
BaseTen$1.40$4.40$0.14$0.52100.0%fp4
Mistral$1.40$4.40$0.14$0.5299.6%nvfp4
Morph$1.01$4.71$0.20$0.53100.0%fp8
Alibaba$1.19$3.74$0.24$0.54100.0%—
Doubleword$1.36$5.44$0.14$0.56100.0%nvfp4
Friendli$1.26$3.96$0.23$0.56100.0%—
Mistral$1.54$4.84$0.15$0.5899.3%nvfp4
AtlasCloud$1.40$4.40$0.26$0.62100.0%fp8
Baidu$1.40$4.40$0.26$0.62—fp8
Crusoe$1.40$4.40$0.26$0.62100.0%fp4
Fireworks$1.40$4.40$0.26$0.62100.0%—
Modal$1.40$4.40$0.26$0.6299.9%—
Nebius$1.40$4.40$0.26$0.62100.0%fp4
Parasail$1.40$4.40$0.26$0.62100.0%fp4
Together$1.40$4.40$0.26$0.6299.7%—
Venice$1.40$4.40$0.26$0.6299.4%—
Z.AI$1.40$4.40$0.26$0.62100.0%fp8
Makora$1.50$4.40$0.26$0.6399.0%fp4
Amazon Bedrock$1.68$5.28$0.31$0.74100.0%—
Alibaba$1.68$5.28$0.34$0.76100.0%—
BaseTen$2.10$6.60$0.21$0.78100.0%fp8
BaseTen$2.10$6.60$0.21$0.78100.0%fp8
Parasail$2.10$6.60$0.26$0.82100.0%fp4
Fireworks$2.10$6.60$0.39$0.9399.7%—
ModelRun$2.10$6.60$0.40$0.94100.0%fp4