| Plan | Amount | Frequency | Channel(s) |
|---|---|---|---|
|
Two machines for reproducible benchmarks
Everything runs on one MacBook Pro M2 Max (64 GB) that is also the development machine. Reviewers of the scaling paper asked for a second host, x86/Linux, CUDA and MPS timings and repeats at 100k and 1M records; one BrainCore pass is 80.9 million spike events and blocks the laptop for hours. Plan: a MacBook Pro 16 M5 Max 128 GB / 8 TB and an x86 Linux workstation with an RTX 5090. Deliverables within three months: the requested measurements, published on Zenodo under CC BY. |
EUR 16,000 | One-Time | |
|
Benchmark evaluation runs
API cost of the answering and judge models for repeated benchmark runs (three runs per configuration on LoCoMo and LongMemEval-S, one fixed judge across all published configurations). Every run is published with its config, per-question answers and verdicts. |
EUR 300 | Monthly | |
|
Any amount
Goes to the two plans above in that order. |
EUR 0 | Other |
| Channel | Type | Address |
|---|---|---|
| Github-Sponsors |
Payment-Provider
GitHub Sponsors, one-time or monthly. |
Visit |
| Bank |
Bank
Bank transfer (EUR or USD) to the maintainer's account in Serbia; details on request by email. |
— |