# InferenceNet

HQ US · serves 7 tracked open-weights models. status page · privacy policy

models
7
verified
—
of overall top 10
3
cheapest on
1 models
vs best $
1.63×
ctx (median)
1024k
uptime
100.0%

# models

model rank $/M in $/M out $/M cache src ✓
glm-5.3 #1 $0.08 $3.60 $0.08 openrouter —
kimi-k3 #2 $0.24 $10.80 $0.21 openrouter —
glm-5.3-flash #3 $0.12 $0.39 $0.03 openrouter —
deepseek-v4.1-flash — $0.10 $0.65 $0.01 openrouter —
deepseek-v4.1-flash-1017 — $0.08 $0.50 $0.01 openrouter —
minimax-m3.1-flash — $0.10 $0.55 $0.01 openrouter —
step-5-preview — $1.00 $2.70 $0.05 openrouter —

# recent changes

when (UTC)modelchange
Oct 25 09:53 deepseek-v4.1-flash $/M output $0.71 → $0.65 ↓ 9.8%
Oct 25 09:53 glm-5.3 $/M output $4.00 → $3.60 ↓ 10.0%

all providers · machine-readable: providers.json