What is Inworld Realtime Inference?
Inworld's hosted open-model LLM API for consumer apps, priced up to 50% below public rates.
Who is Inworld Realtime Inference best for?
Inworld Realtime Inference is best for consumer app teams with large LLM bills who want fast open models.
High-volume consumer apps
Inworld says its hosted open models run up to 50% below the public third-party rate, which matters when most of an app's users never pay.
Mixing open and closed models
The Router sends each request to an Inworld-hosted model or a provider such as Anthropic from one config, with one bill and no routing markup.
Voice apps
The same account and credits cover Inworld's text-to-speech and speech-to-text, so volume discounts apply across the voice stack.
Very large workloads
When volume justifies it, you can move from per-token billing to dedicated GPUs from $5 per GPU-hour.
How much does Inworld Realtime Inference cost?
Inworld Realtime Inference is billed per token on Inworld's plans, from a free pay-as-you-go On-Demand tier to monthly plans with $25, $100 or $300 in credits, and its hosted models start at $0.07 per million input tokens.
- Hosted LLMs billed at listed per-token rates
- 220+ models through the Router with no markup
- Up to 70 minutes of text-to-speech included
- $25 in credits a month for TTS, STT and LLMs
- Up to 33% off TTS and STT rates
- Workspaces and team management
- $100 in credits a month
- Up to 40% off TTS and STT rates
- Higher concurrency limits
- $300 in credits a month
- Up to 47% off TTS and STT rates
- Priority email support
Inworld-hosted models are priced per million tokens, from $0.07 input and $0.34 output for Gemma 4 26B A4B to $1.40 and $4.40 for GLM 5.3, and two DeepSeek V4 Flash variants are listed as free. Growth is $1,500 a month and Enterprise is custom; yearly billing gives 2 months free, and dedicated GPUs start at $5 per GPU-hour. Full pricing on inworld.ai. We checked these prices on .
How do readers rate Inworld Realtime Inference?
No one has rated Inworld Realtime Inference yet. Be the first.
Sign in to rate. One rating per account.
What are the best Inworld Realtime Inference alternatives?
Three tools on Recaply do Inworld Realtime Inference's job.
Questions about Inworld Realtime Inference
What is Inworld Realtime Inference?
Inworld's hosted open-model LLM API for consumer apps, priced up to 50% below public rates.
Who is Inworld Realtime Inference for?
Inworld Realtime Inference is best for consumer app teams with large LLM bills who want fast open models.
Is Inworld Realtime Inference free?
Yes. Inworld Realtime Inference has a free On-Demand plan: hosted LLMs billed at listed per-token rates.
How much does Inworld Realtime Inference cost?
Inworld Realtime Inference's paid plans are Creator at $25 per month, Builder at $100 per month and Developer at $300 per month.
What are the best Inworld Realtime Inference alternatives?
The closest alternatives on Recaply are Together AI, Featherless and Avian.
Where can I try Inworld Realtime Inference?
On its official website, inworld.ai.
Can Inworld host a model that is not on its list?
Yes. Inworld says that if the model you need is not in its lineup, its inference team will optimize it and serve it on the same stack with the same latency, reliability and cost profile. You arrange this through Inworld's sales team.







