SShogo Bench

GLM-5.3-Flash

High-speed GLM-5.3-Flash serving option for coding and agent workflows

Shogo Agent Index

Supported#38 of 92
48.3range 36.5–60 · 27% of weight measured
Agentic coding
54
43% measured
Long-horizon autonomy
57
20% measured
Knowledge-work agents
40
32% measured
Computer use
34
estimated
Security & context learning
47
estimated

Price

Input
$0.30/M
Output
$1.05/M
Cache read
$0.085/M
3:1 blend
$0.49/M
Agent-effective
$0.16/M

Cheapest host for agents: Darkbloom at $0.030/M.

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

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verifiedest.8%predicted 77
Terminal-Bench 2.0est.8%predicted 68
DeepSWE63.4%6%mini-swe-agent · $0.48/task
FrontierSWE18.1%5%proximus · $11.03/task
CursorBench36.8%4%source · $0.39/task
GSO-Benchest.4%predicted 26
METR Time Horizonsest.10%predicted 74
Vending-Bench 2est.6%predicted 43
PostTrainBench36.8%4%Claude Code
APEX-Agents52.8%8%source
GDPvalest.7%predicted 45
Remote Labor Indexest.5%predicted 3
DeepResearch Benchest.5%predicted 48
OSWorld 2.0est.6%predicted 8
OSWorldest.4%predicted 73
Cybenchest.5%predicted 72
CL-benchest.5%predicted 21

Other benchmarks

BenchmarkCategoryScoreSource
ARC-AGIReasoning91.0%source
ARC-AGI-2Reasoning65.8%source
Chess PuzzlesReasoning21.0%source
CritPtReasoning15.4%source
FrontierCodeCoding31.8%devin-cli
FrontierMath Tiers 1-3Math55.8%source
FrontierMath Tier 4Math17.1%source
GPQA DiamondReasoning90.2%source
OTIS Mock AIMEMath93.9%source
ProofBenchMath21.0%source
SciCodeCoding51.6%source
WebDev ArenaCoding1607source

API hosts

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

HostInputOutputCache readAgent $/MUptime 24hQuant
Darkbloom$0.065$0.18$0.012$0.03095.3%fp4
GMICloud$0.075$0.25$0.015$0.03099.2%fp8
Relace$0.040$0.078$0.030$0.030100.0%fp4
Tenstorrent$0.075$0.25$0.015$0.03099.8%fp8
Near AI$0.080$0.50$0.012$0.04099.9%fp8
Novita$0.084$0.28$0.017$0.040100.0%fp8
Particle AI$0.065$0.25$0.020$0.040100.0%nvfp4
StreamLake$0.088$0.29$0.018$0.04091.7%fp8
CoreWeave$0.12$0.38$0.020$0.05099.9%nvfp4
OpenInference$0.050$0.45$0.025$0.05099.9%fp4
Wafer$0.042$0.18$0.041$0.050100.0%nvfp4
DekaLLM$0.10$0.50$0.030$0.06099.6%—
DigitalOcean$0.11$0.38$0.030$0.06099.9%—
InferenceNet$0.11$0.35$0.035$0.060100.0%fp4
InferenceNet$0.11$0.35$0.035$0.060100.0%fp4
Makora$0.14$0.24$0.030$0.06099.6%fp4
Reka$0.12$0.30$0.030$0.060100.0%—
River$0.12$0.40$0.024$0.06099.7%—
SiliconFlow$0.10$0.40$0.030$0.06092.6%fp8
Baidu$0.15$0.50$0.030$0.07087.0%fp8
BaseTen$0.15$0.50$0.030$0.070100.0%fp8
Cohere$0.15$0.50$0.030$0.070100.0%nvfp4
Crusoe$0.15$0.50$0.030$0.070100.0%fp4
DeepInfra$0.15$0.50$0.030$0.070100.0%fp4
Fireworks$0.15$0.50$0.030$0.070100.0%—
Friendli$0.15$0.50$0.030$0.07099.8%—
KT$0.15$0.50$0.030$0.07099.6%—
Mistral$0.15$0.50$0.030$0.070100.0%—
Modal$0.15$0.50$0.030$0.070100.0%fp4
Nebius$0.15$0.50$0.030$0.070100.0%fp4
Parasail$0.15$0.50$0.030$0.07099.9%fp8
Particle AI$0.13$0.50$0.040$0.070100.0%nvfp4
Sail Research$0.17$0.63$0.022$0.070100.0%fp4
Sail Research$0.17$0.63$0.022$0.070100.0%fp4
Together$0.15$0.50$0.030$0.070100.0%—
Venice$0.15$0.50$0.030$0.070100.0%—
Wafer$0.065$0.24$0.064$0.070100.0%nvfp4
Z.AI$0.15$0.50$0.030$0.070100.0%fp8
Inceptron$0.15$0.50$0.040$0.08099.9%fp8
Morph$0.12$0.36$0.062$0.08099.6%fp8
Parasail$0.19$0.63$0.037$0.09099.8%fp4
Fireworks$0.23$0.75$0.045$0.10100.0%—