inferenstack — the market read on AI inference · last observation 2026-09-20. See the boards
Price spread

Qwen3.8 2.4T A95B

alibaba· 2026-08-12· 262K ctx
open weightstool callreasoningstructured quant: unknown

spread (output)

1.3×

$5.95 – $7.50 · 18 providers

18 providers priced · 1 omitted (stale)
ProviderInputOutputCache rdBlended
IteraCompute
relay router
$1.95$5.95$0.200$1.20
Cortecs
relay router
$2.50$6.00$0.625$1.63
Deep Infra
open weights host
$2.00$6.00$0.200$1.22
Eden AI
relay router
$2.00$6.00$0.250$1.45
Fireworks AI
relay router
$2.00$6.00$0.250$1.25
Charm Hyper
relay router
$2.00$6.00$0.250$1.25
AIHubMix
unknown
$2.00$6.00$0.500$1.41
RunInfra
relay router
$2.00$6.00$0.200$1.22
TensorX
relay router
$2.50$6.00$0.630$1.63
Vercel AI Gateway
unknown
$2.00$6.00$0.250$1.25
Kilo Gateway
relay router
$2.00$6.00$0.250$1.45
Modal
relay router
$2.00$6.00$0.250$1.25
NanoGPTstale
relay router
$2.00$6.00$0.250$1.45
OpenRouter
unknown
$2.00$6.00$0.250$1.25
Requesty
relay router
$2.00$6.00$0.200$1.22
Requesty
relay router
$2.50$6.00$0.630$1.63
Hugging Faceoutlier
relay router
$2.50$6.25—$2.84*
Merge Gatewayoutlier
unknown
$2.50$6.25$0.500$1.57
Venice AIoutlier
unknown
$2.50$7.50$0.313$1.56

Data: models.dev (MIT), snapshot 2026-09-20. Prices are vendor-published; verify before purchase. Methodology.