Home › Blog › Cost Calculator

LLM API Cost Calculator (2026)

Estimate your monthly spend across DeepSeek, GLM, Kimi and Qwen — with the exact pricing Bridge uses in production billing. Move the sliders; the table updates instantly.

Requests per day

5,000

Prompt tokens per request

2,000

Completion tokens per request

500

Cache-hit rate (share of prompt served from cache, 10× cheaper)

50%

Estimated monthly cost

Model	Input / output (Bridge, per 1M)	Monthly cost

Prices: DeepSeek V4 Pro $0.25/$0.65 (cache $0.025) · GLM-5.2 $0.40/$1.20 (cache $0.04) · Kimi K3 $1.50/$3.50 (cache $0.15) · Qwen2.5 72B $0.25/$0.75. These mirror the live price table; models without a cached-input tier are computed without one.

How the math works

monthly = requests × days × [ prompt × (miss_rate × input + hit_rate × cached) + completion × output ] ÷ 1M

Three levers move the number most: cache-hit rate (10× on cached input), completion length (output prices are 2–3× input), and model choice — the same 10M monthly tokens cost ~4× more on Kimi K3 than on DeepSeek V4 Pro. See how the four compare in our full model comparison.

Get $1 free credit → meter real usage