NeuroXess-N1
NeuroXess · Undisclosed (neural encoder + vision tower, MoE)
The only Chinese multimodal model that takes neural signals as an input modality
A multimodal model from an implantable BCI company: EEG/fNIRS temporal signals, images and text in one discrete token space, aimed at neuro-rehabilitation assessment and embodied interaction. Composite 77.2, ranked 30th, and one of two grade-C models here (the other is Hikvision Guanlan 3) — also the only model here with no runner-up: on neural-signal decoding, it is the sole entry.
Last updated Oct 18, 2026Editorial scoring
Bottom line
Eight-dimension breakdown
editorial mapping · 0–100- NeuroXess-N177.2
Weights: perception 20 / documents 18 / charts 14 / video 14 / agents 14 / generation 6 / Chinese 6 / ecosystem 8. Definitions and sources on the methodology page.
Public benchmarks
as reported, not converted| Benchmark / metric | Value | Note |
|---|---|---|
| 神经信号语义解码 Top-1 | 62% | 10-word closed set, official |
| EEG-图像跨模态检索 mAP | 71.3 | Official, not independently reproduced |
| 部署形态 | 院内私有化 | Project licensing; weights stay on-site |
Public scores come from different versions and prompts and are not interchangeable with our composite; they are one evidence input.
Scenario mapping
five stress tests, one scale| Scenario | Score | Observation |
|---|---|---|
| Chinese receipt OCR | 80 · Usable on standard receipts, not a strength | |
| Chart data extraction | 74 · Standard charts | |
| 30-min video QA | 68 · Summary level only | |
| Mobile GUI control | 70 · Not a target scenario | |
| Neural-signal decoding | 88 · The only entry on this board |
Strengths and limitations
Strengths
- End-to-end neural-signal-to-semantics decoding: EEG/fNIRS shares a token space with images and text, not a bolted-on classifier.
- In-hospital on-prem delivery experience: the data-compliance path is designed for medical settings, which general API vendors cannot offer.
- Chinese context 88: markedly better priors on medical records and rehabilitation scales than its score peers.
Limitations
- Grade C: every number is vendor-reported, with zero third-party reproduction.
- Ecosystem 62 is the lowest here: no public API, no SDK docs, no rate-limit documentation.
- Weak across general dimensions: generation 68, video 70 — outside neural and embodied scenarios it is unusable.
Price and real cost
List price
—N/A
No public API; project licensing. For 1,000 images at 1,024px, ~1,100 image tokens each and 300 output tokens, the monthly cost is about ¥0.0 (assuming input at 25% of output price).
Use your volumeAbout NeuroXess-N1
Is NeuroXess-N1 open source?
Licence: Closed · project licensing. No public weights; access is via vendor API or an on-prem licence.
How much does NeuroXess-N1 cost?
No public API; project licensing. It ranks #30 with a composite of 77.2. Real bills depend on resolution, tiling, caching and volume discounts.
What is NeuroXess-N1 best for?
Neuro-rehabilitation assessment and stroke follow-up research; Embodied interaction and neural-interface prototypes; In-hospital projects where data compliance leads。
Why does NeuroXess-N1 score 77.2?
The composite weights eight dimensions: perception 20%, documents 18%, charts 14%, video 14%, agents 14%, generation 6%, Chinese 6%, ecosystem 8%. Its strongest dimension is Chinese context & culture; its weakest is Ecosystem, cost & availability.
Personal ranking · Not an official score
Every ranking, score and comparison on this page is an independent personal ranking, not an official score, and represents no vendor, institution or regulator. Scores are a subjective mapping of public evidence under our own weights — change the weights and the order changes. Disagree, or spotted a factual error? Contact the site owner: jizhi2024@foxmail.com — we check each report and log corrections in the changelog.