Model comparisons

InferIndex · https://inferindex.dev/en/compare/deepseek/deepseek-v3.2-exp/vs/tencent/hy3-preview

DeepSeek V3.2 Exp costs from $0.29 per million tokens and Hy3 preview from $0.285 (at a 3:1 input:output mix): the cheapest price for each; prices checked on 31 October 2026.

Each model's lowest blended price, per 1M tokens, across every provider we track.

DeepSeek V3.2 Exp vs Hy3 preview: API prices

Model	Blended$/1MLowest blended price, $ per 1M tokens. Green background: the most favourable value of each column.	Input$/1MInput price of the cheapest offer, per 1M tokens	Output$/1MOutput price of the cheapest offer, per 1M tokens	No training$/1MThe provider's published policy says the text you send is not used to train models. We only show what is published; we can't check it ourselves.	Monthly cost$For 2,000 input tokens and 500 output tokens per request, 1,000 requests a day, at the cheapest offer for that workload. Change the workload above.	Offers	Providers	ContextThe maximum amount of text, counted in tokens, the model can take in and write in a single request. Past this limit, the request is refused or cut.	Observation
Hy3 preview, Tencentat GMICloud via OpenRouter	$0.285, most favourable value of the column	$0.18, most favourable value of the column	$0.60	–	$19.80, most favourable value of the columnat GMICloud via OpenRouter	5	5	262k	CheapestCheapest. Lowest blended price, $ per 1M tokens
DeepSeek V3.2 Exp, DeepSeekat AtlasCloud	$0.29Most expensive offer · $0.315. The highest blended price among comparable offers: one price per provider (its cheapest route), standard tier, current prices only. Priority tiers and prices converted from points are not counted.	$0.26	$0.38, most favourable value of the column	$0.315at NanoGPT	$21.30at AtlasCloud	8	6, most favourable value of the column	164k

Green background: the most favourable value of each column.