Roadmap
What works today, what is being built, and what is not on the table. There are no dates here: a feature gets a date once it has run on real GPUs with real payments.
- Live
Live now
- GPU catalog with prices read from the Nosana markets on Solana and each market's queue as it stands.
- Provider waitlist.
- Early access
Now in early access: AI Studio
- Four open-weight models on GPUs with 24 GB of memory: gpt-oss-20b, Qwen3.5 9B, Gemma 4 26B and SDXL. FLUX.2 klein and Whisper follow after a test on real nodes.
- Payment from your own wallet, straight to the Nosana program, with the fee shown before you sign.
- Private endpoint and a playground for each running model. Stop any time; unused time comes back.
- Tested with small paid jobs on mainnet: deploy, chat, added time, and stop with refund.
- In development
Next: Render
- Blender Cycles renders split across several GPUs, with progress per frame and downloads as frames or MP4.
- Scene checks before you pay, and a probe render for an honest quote.
- Planned
Later, not scheduled
- Gridora's own provider program, with an open-source benchmark tool for hosts.
- Jobs on confidential-computing hardware for workloads that need it, once it can be verified end to end.
- Card payments through Nosana's own accounts, for people without a crypto wallet.