LLM Cost Calculator — Real Monthly Cost, Not Just Rate

List price lied to you

These models move the most once verbosity is priced in — a model that talks more costs more, regardless of its list rate.

• GLM 4.7 (Cerebras) is priced #44 by list rate but #61 once its 7.53× verbosity is billed — $1,273 list vs $2,530 effective.
• GLM-5.2 is priced #38 by list rate but #54 once its 3.87× verbosity is billed — $980 list vs $1,864 effective.
• DeepSeek V4 Pro is priced #37 by list rate but #50 once its 3.30× verbosity is billed — $911 list vs $1,548 effective.
• Muse Spark 1.3 Contributor is priced #5 by list rate but #17 once its 12.95× verbosity is billed — $62.00 list vs $229 effective.
• Gemini 3.1 Pro is priced #55 by list rate but #48 once its 0.63× verbosity is billed — $1,800 list vs $1,489 effective.
• Mistral Medium 3 is priced #43 by list rate but #37 once its 0.84× verbosity is billed — $1,245 list vs $1,161 effective.

The formula, published

outputTokensBilled = outputTokens × (verbosityIndex ?? 1)
inputCost  = inputPerM  × inputTokens        / 1e6
outputCost = outputPerM × outputTokensBilled / 1e6
effective  = (inputCost + outputCost) × callsPerMonth

with caching: inputCost × (1 − cacheableInputPct × 0.9)
with batching: (inputCost + outputCost) × (1 − batchDiscountPct/100)

up to 90% cache-read saving is the largest sourced read discount in the provider table; provider-specific estimates use that provider's multiplier and account for cache writes. The free-form estimator assumes 30% of input is cache-eligible; pick a workload preset below for a shape-specific assumption instead.

Coverage: 20 of 76 priced models have a measured verbosity index (2+ graded runs). The rest render with a verbosity of — and are shown at unadjusted list price.

Cost by workload shape