# Inferenceapis

serves 14 tracked open-weights models.

models
14
verified
—
of overall top 10
2
cheapest on
—
vs best $
2.96×
ctx (median)
1048k

# models

model rank $/M in $/M out src ✓
deepseek-v4-pro #1 $1.72 $5.15 litellm —
minimax-m3 #2 $0.39 $1.56 litellm —
deepseek-v3 — $0.19 $0.38 litellm —
deepseek-v4-flash — $0.19 $0.38 litellm —
deepseek-v4-flash-0731 — $0.19 $0.38 litellm —
deepseek-v4-pro-0813 — $1.72 $5.15 litellm —
deepseek-v4.1-flash — $0.40 $1.60 litellm —
glm-5.3 — $1.82 $5.72 litellm —
glm-5.3-flash — $0.20 $0.66 litellm —
gpt-oss-120b — $0.20 $0.80 litellm —
gpt-oss-20b — $0.04 $0.19 litellm —
kimi-k2.7-code — $0.89 $4.42 litellm —
llama-3.3-70b-instruct — $0.20 $0.50 litellm —
qwen3-vl-235b-a22b-instruct — $0.26 $1.15 litellm —

all providers · machine-readable: providers.json