Set the token volumes, and every model is priced against them at once — the Apertis rate beside the market list rate, both dated.
Monthly agent workload
K
100M tokens per month
K
40M tokens per month
Estimated monthly cost, every model at once
Add Credits purchases include bonus credit: +2% from $500, +5% from $2,000 and +8% from $10,000. The Apertis column is the price you pay, without bonus credit. See the schedule.
Estimated monthly cost for 100M input and 40M output tokens, cheapest first.
Model
Apertis
Market list
Difference
Nemotron 3.5 Super VL
$26.00$0.10 in · $0.40 out / 1M
$40.00$0.20 in · $0.50 out / 1M
−$14.0035%
DeepSeek V4.1 Flash 1017
$52.00$0.20 in · $0.80 out / 1M
$78.00$0.30 in · $1.20 out / 1M
−$26.0033%
Muse Spark Open
$295.00$1.25 in · $4.25 out / 1M
$295.00$1.25 in · $4.25 out / 1M
Same
GPT-6.1 Terra
$300.00$1.00 in · $5.00 out / 1M
$300.00$1.00 in · $5.00 out / 1M
Same
Gemini 4 Carbon
$600.00$2.00 in · $10.00 out / 1M
$600.00$2.00 in · $10.00 out / 1M
Same
Kimi K3.1 Pro
$600.00$2.00 in · $10.00 out / 1M
$600.00$2.00 in · $10.00 out / 1M
Same
Claude Fable 5.2
$2.4K$8.00 in · $40.00 out / 1M
$2.4K$8.00 in · $40.00 out / 1M
Same
Apertis figures come from the Apertis models catalog. Market list figures come from Artificial Analysis. Both read 2026-11-17. A month's purchase sets its bonus credit tier, and you always pay the amount you choose.
Reference estimate only. Your Apertis bill depends on the current access option, included usage, overflow terms, routed provider, and actual cache behavior. This is not a quote or a guaranteed savings calculation.
Open models next to Claude
List prices for each vendor's flagship coding model, priced on the workload above.