Data

When inference host prices differ, and when they do not

· Prices verified November 16, 2026 to November 18, 2026

Of the 9 models served by two or more hosts in this dataset, 6 cost practically the same wherever they run: every host is within 1.15× of the cheapest. Another 3 show a real spread, the widest being DeepSeek V4 Flash 0731 at 7.3×. For the first group, the per-token price does not separate the hosts. For the second, the most expensive host charges a multiple of the cheapest.

Every price below is a verified host price with a source link; 36 offerings of 14 models across 7 hosts. How this is checked

Most, but not all

Every model here is one that several companies can serve, which is what makes a comparison possible at all; a proprietary model has exactly one seller and belongs in the pricing table, not here. Among these, for some the market has converged and for others it has not. The dataset behind the provider comparison draws the line at 1.15×: if the most expensive host charges no more than 1.15× the cheapest per input token, the per-token price does not separate the hosts, and we call that "the same price". Today that line splits the 9 multi-host models into 6 that cost the same and 3 that do not.

Where every host charges about the same

6 models: every host within 1.15× of the cheapest on input.

Model Hosts Input Output Spread
DeepSeek V4 Pro 0813 Together AI · OpenRouter · Baseten $1.32 $3.96 none
GLM-5.3 Together AI · Fireworks AI · OpenRouter · Novita AI · Baseten $1.40 $4.40 none
GLM-5.3 Flash OpenRouter · Baseten $0.15 $0.50 none
Kimi K3 DeepInfra · Together AI · Fireworks AI · OpenRouter · Novita AI · Baseten $2.85–$3.00 $14.25–$15.00 1.1×
MiniMax M3 Fireworks AI · OpenRouter $0.30 $1.20 none
Muse Spark Open Together AI · Fireworks AI · Baseten $1.25 $4.25 none
DeepSeek V4 Pro 0813: OpenRouter's $1.32 / $3.96 is the maker's list rate, which applies only during the maker's peak hours (Monday to Friday 01:00-04:00 and 06:00-10:00 UTC: 35 of the 168 hours in a week, 21%); the rest of the time that host, like the maker's own API, charges half, $0.66 / $1.98. Together AI and Baseten charge $1.32 / $3.96 at all hours.

DeepSeek V4 Pro 0813, GLM-5.3, GLM-5.3 Flash, MiniMax M3 and Muse Spark Open cost exactly the same on every host listed. The widest gap among them is 1.1× on input.

Where the hosts' prices differ

3 models, widest spread first. Cheapest and most expensive host for identical weights.

Model Cheapest host Most expensive host Spread
DeepSeek V4 Flash 0731 DeepInfra $0.06 in · $0.18 out Novita AI $0.44 in · $1.32 out 7.3×
gpt-oss-120b Novita AI $0.05 in · $0.25 out Fireworks AI, Groq and Together AI $0.15 in · $0.60 out 3.0×
Llama 3.3 70B Instruct Novita AI serves it with a 12K-token context; DeepInfra with 131K. DeepInfra $0.10 in · $0.32 out Novita AI $0.135 in · $0.40 out 1.4×

The spread runs from 1.4× on Llama 3.3 70B Instruct to 7.3× on DeepSeek V4 Flash 0731: at the top end, the most expensive host charges $0.44 per million input tokens for weights another host serves at $0.06. On these models the most expensive host charges more than 1.15× the cheapest. Full host-by-host tables, with every price linked to its source, are on the provider comparison.

Three things to keep in mind

A maker's own list price anchors the hosts. Where the maker serves the model itself with a published price, the other hosts sit at or near it (DeepSeek V4 Pro 0813, GLM-5.3, GLM-5.3 Flash, Kimi K3 and MiniMax M3). All 3 models with a real spread (DeepSeek V4 Flash 0731, gpt-oss-120b and Llama 3.3 70B Instruct) lack that anchor: no host is the maker, so each sets its own price. The reverse does not hold: 1 model converges with no maker among the hosts (Muse Spark Open), so the anchor helps, but it is not the only thing that keeps hosts in line.

Spread is on input price. The ratio compares the most and least expensive host per million input tokens, because that is where hosts differ most and what most workloads are dominated by. Output prices usually move with them; the tables show both.

OpenRouter is a router, not a host. Where OpenRouter appears in the tables, the price is the model maker's own list price on OpenRouter's providers page, without promotional discounts. With :floor or sort: "price", OpenRouter routes to the cheapest reseller at that moment, often below that list price; that price changes on its own as promotions end and hosts come and go. A default request is spread across several hosts and pays something in between. The mechanics, with evidence, are in OpenRouter pricing: the number in the catalog isn't what you pay. "No list price" on this page means the maker is not among OpenRouter's hosts for that model. It is a different thing from "no published price" on the pricing table (a maker that sells the model but publishes no per-token price for it: today Cohere for 4 current models) and from "no data" (a price we could not verify).

A further 7 models in the dataset have a single host; for 2 of them there is no list price to show. There is nothing to compare there: each of these models is available from one host only.

Verification

All figures on this page are computed at build time from the same dataset and the same rules as the provider comparison, so the two pages cannot disagree. An offering is included only if it carries a verification date and a link to the host's official pricing page; verification dates range from November 16, 2026 to November 18, 2026. The 1.15× threshold is a fixed editorial cut, not a measurement; move it and the split moves with it. Prices are USD per million tokens. Nothing is estimated.