Qwen3 14B pricing.
prices as of · 19 providers · 20 prices
Alibaba Cloud's Qwen3 14B holds 131k tokens of context and has a reasoning mode. Open weights, so anyone may host it, which is why the prices below differ so much.
Qwen3 14B costs $0.35 in and $1.40 out per million tokens from Alibaba Cloud; the cheapest price we can explain is $0.41 at Nscale, checked 16 Sept 2026.
Alibaba Cloud list price
$2.45/M tokens
$0.35 in · $1.40 out
read · 31% below the market median
Cheapest price we can explain
$0.41/M tokens
$0.07 in · $0.20 out · 83% under list
Nscaleread
What it costs for your work
Alibaba Cloud list price
$0.294a month
cheapest we can explain · Nscale
$0.054a month
difference
$0.24saved
A page is about 600 tokens in and 60 out, a support reply 200 in and 300 out, an email 150 in and 250 out. Nothing else is counted: no cache discount, no batch rate, no free tier.
Run it yourself
every plan that fits Qwen3 14B →Alibaba Cloud publishes the weights, so you can run Qwen3 14B on a server you rent by the month. It needs about 11.8 GB of RAM on a CPU, and the cheapest plan we track with that much and a disk included is HOSTKEY vm.v2-medium at $16.15/mo, streaming an estimated 2.1–4.3 tok/s.
Every provider selling Qwen3 14B
Four groups, decided by the price and by what the provider publishes about it. Prices are per million tokens, and the share is that row against Alibaba Cloud's price.
The maker's list price
1 priceWhat the lab that trained the model charges on its own API.
Alibaba CloudThe price Alibaba Cloud charges for the model it trained. Every row below is measured against it.
$0.35 / $1.40in / out per M
1.00×of list price
Cheaper, and we know why
17 pricesBelow list, with a published reason: hosting the open weights, a regional price list, credits, a stated discount.
- Nscale
Hosts the open weights on its own servers, so the price is its own, not a resale. Precision not disclosed.
41k context hereNscale ↗$0.07 / $0.20in / out per M
0.17×of list price
TrustedRouterRoutes to a host serving the open weights; precision not disclosed.
41k context hereTrustedRouter ↗$0.07 / $0.21in / out per M
0.18×of list price
$0.08 / $0.24in / out per M
0.20×of list price
Nebius Token FactoryHosts the open weights on its own servers, so the price is its own, not a resale. Precision not disclosed.
33k context hereNebius Token Factory ↗$0.08 / $0.24in / out per M
0.20×of list price
SiliconFlow (China)SiliconFlow's China-region catalog, priced in the domestic market.
131k context hereSiliconFlow (China) ↗$0.07 / $0.28in / out per M
0.20×of list price
SiliconFlowHosts the open weights on its own servers, so the price is its own, not a resale. Precision not disclosed.
131k context hereSiliconFlow ↗$0.07 / $0.28in / out per M
0.20×of list price
NextBitServes the open weights at int4, a lower precision than the maker's, so not quite the same product.
41k context hereint4 precision$0.10 / $0.22in / out per M
0.21×of list price
Nous PortalResells OpenRouter's catalog at 0.80× the listed price, funded by a subscription whose credits carry a 10% bonus.
$0.10 / $0.22in / out per M
0.21×of list price
$0.12 / $0.24in / out per M
0.24×of list price
DeepInfraHosts the open weights on its own servers, so the price is its own, not a resale. Precision not disclosed.
41k context hereDeepInfra ↗$0.12 / $0.24in / out per M
0.24×of list price
- DeepInfra
Serves the open weights at fp8, a lower precision than the maker's, so not quite the same product.
$0.12 / $0.24in / out per M
0.24×of list price
FastRouterRoutes to a host serving the open weights; precision not disclosed.
41k context hereFastRouter ↗$0.12 / $0.24in / out per M
0.24×of list price
Eden AIAlibaba's China-region price list, which is lower than the international one. A region, not a deal.
41k context hereEden AI ↗$0.12 / $0.24in / out per M
0.24×of list price
Fireworks AIHosts the open weights on its own servers, so the price is its own, not a resale. Precision not disclosed.
41k context hereFireworks AI ↗$0.20 / $0.20in / out per M
0.33×of list price
Alibaba Cloud (China)Alibaba's China-region Model Studio, which prices the Qwen models below the international list. A region, not a deal.
$0.14 / $0.57in / out per M
0.41×of list price
Kilo GatewayRoutes to a host serving the open weights; precision not disclosed.
41k context hereKilo Gateway ↗$0.23 / $0.91in / out per M
0.65×of list price
Inference.netHosts the open weights on its own servers, so the price is its own, not a resale. Precision not disclosed.
$0.23 / $0.91in / out per M
0.65×of list price
Cheaper, and nobody says why
2 prices · never rankedBelow list on a closed model with nothing to account for it. Shown, never ranked, never a pick.
Weights & Biases0.14× the list price, from one source and far below what hosting the weights costs. Listed until a second source agrees.
33k context here$0.05 / $0.22in / out per M
0.15×of list price
FeatherlessWe could not read this provider's prices reliably, so the row is listed but never ranked.
33k context hereFeatherless ↗$0.12 / $0.24in / out per M
0.24×of list price
Frequently asked
- Who sells Qwen3 14B cheapest?
- Nscale, at $0.07 in and $0.20 out per million tokens, read on 16 Sept 2026. Hosts the open weights on its own servers, so the price is its own, not a resale. Precision not disclosed.
- Is Qwen3 14B open weights?
- Yes. Alibaba Cloud publishes the weights, so anyone may host Qwen3 14B, which is why 19 providers sell it and why their prices differ so much.
- How much context does Qwen3 14B hold?
- 131,072 tokens in one call, which is about 131k, and it can call your own functions.
- How does Qwen3 14B compare with the rest of the market?
- At $2.45 per million tokens, Alibaba Cloud's price is 31% below the $3.55 median across the 337 text models we track.