Sunk Cost sunkcost.ai Data checked 2026-09-03

Tencent Hy3 vs Sage S2.1 Flash

Tencent Hy3 scores 26 on the intelligence index, and Sage S2.1 Flash has not been placed on it. None of the 37 machines here still sold at a published price runs either model at 32k of context.

Tencent Hy3Sage S2.1 Flash
Intelligence index26not placed
ClassHaiku-classnot yet placed
Weights182 GB194 GB
Needs at 32k193 GB194 GB
QuantisationQ4_K_MQ4_K_M
Parameters295B (21B active)321.3B (18B active)
Max context256k1M
API price per 1M$0.0825 in / $0.33 out$0.0399 in / $0.0784 outpriced as GLM-5.3-Flash
LicenceApache 2.0MIT
Machines here that run it0 of 370 of 37
Cheapest machine that runs itnone listednone listed

Run Tencent Hy3 in the calculator · or Sage S2.1 Flash in the calculator

Memory is not what separates them

Tencent Hy3 needs 193 GB of memory at 32k of context and Sage S2.1 Flash needs 194 GB. Every machine priced here that runs one runs the other, at 32k of context. So the choice between them is what each is good at, how fast it runs and what the same work costs on an API, not what you have to buy to hold it.

Tencent Hy3 is also head to head with Inkling Small above it on the leaderboard and Ling 3.0 flash below it. Two more models need much the same memory: GLM-5.3-Flash and GLM-5.3-SLIM.

The assumptions behind both columns

Both columns use the same usage: 500k tokens a day at 15:1 input to output, 32k of context, $0.17 per kWh, and today's API prices held flat. Speeds marked estimated are worked out from memory bandwidth rather than measured, and pay-back scales with them. Where nobody rents an open model by the token, its API prices are the nearest hosted model's, named beside them. Machines are the 37 here with a published price that are still sold. Change any of it in the calculator.

More head to head: what Tencent Hy3 needs · what Sage S2.1 Flash needs · every other match-up · both against the frontier · the quickest pay-back at each level of use