FocsleFocsle
Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/ Fo’c’sle · /fōk-səl/
The platform

Edge AI deployment, charted.

Every model. Every chip. Benchmarked, simulated, and deployable from one source of record. The system of record for the next generation of edge AI products.

/fōk-səl/
Reference models
13K
38 curated reference set
Hardware platforms
17
11 vendors, every major edge SoC
Benchmark cells published
218K
195 in the curated matrix · refreshed weekly
Monthly deployments
1.2M
358K in the reference set
trending now

Models the community is shipping this week

Sorted by week-over-week deployment growth across the curated reference set.

featured benchmark

YOLO27-Edge: every chip, side by side

The data nobody else can publish. Sortable, filterable, with HIL test conditions on every row.

Object detection · COCO 2017 · 1080p input
Object detection
#ModelChipQuantLatency p50ThroughputAcc.Power
1YOLO27-EdgeNNVIDIA Jetson ThorFP166.27 ms160 FPS99.6%122.2 W
2YOLO27-EdgeNNVIDIA Jetson AGX OrinFP169.77 ms102 FPS99.4%49.4 W
3YOLO27-EdgeQSnapdragon 8 Gen 3 NPUINT813.1 ms77 FPS98.0%6.3 W
4YOLO27-EdgeQQualcomm QCS8550INT813.9 ms72 FPS97.9%13.0 W
5YOLO27-EdgeHHailo-10HINT817.0 ms59 FPS98.3%3.3 W
6YOLO27-EdgeNNVIDIA Jetson Orin NanoINT818.1 ms55 FPS97.9%12.2 W
7YOLO27-EdgeAAMD Versal AI Edge VE2302INT818.1 ms55 FPS97.4%10.2 W
8YOLO27-EdgeApple Neural Engine (M4)FP1619.1 ms52 FPS99.9%7.1 W
publishers

Chip vendors, robotics teams, OSS labs — all on the same surface

The two-sided network is working. Vendors publish to be visible on neutral benchmarks. Companies publish to recruit. Labs publish for reach.

Companies & labs deploying production stacks on Fo’c’sle
MERIDIAN AUTONOMYLUMEN ROBOTICSHELIX VISIONAEGIS DEFENSENORTHWIND ENERGYATLAS LOGISTICS
hil simulation

Will it run? Watch it run, before you commit.

The HIL Sim lab pairs every benchmark with an end-to-end run on a real-world scenario — CARLA highway, warehouse AMR aisles, surgical phantoms, drone last-mile. Sustained throughput. In-sim accuracy. Intervention counts. The numbers your customer is going to ask for.

car · 0.94truck · 0.87car · 0.78
CARLA highway · cut-in scenarios
64 FPS · 96.4% acc
car · 0.94truck · 0.87car · 0.78
Tabletop pick-and-place · 200 trials
24 FPS · 88.6% acc
car · 0.94truck · 0.87car · 0.78
Drone last-mile · suburban grid
58 FPS · 92.8% acc
car · 0.94truck · 0.87car · 0.78
PCB defect inspection · line throughput
38 FPS · 97.2% acc
from the blog

Cross-chip benchmark report: Q3 2028

Three transformers caught up to YOLO. Hailo-10 dethroned Coral on the low-power tier. Thor justified its watts on VLA. The full numbers, and the surprises.

Apr 18, 2028 · Fo’c’sle Research · 12 min read
for product teams

Ship edge AI faster, on every chip your customers run.

Private model registries, dedicated HIL lab time, on-prem benchmark publication, ITAR-aligned controls, and a solutions engineer who knows your silicon roster.