Baidu Kunlun P800 Kunlun III
Overview
The Baidu Kunlunxin (昆仑芯) P800 is the 3rd-generation Kunlun (Kunlun III) AI accelerator on the self-developed XPU-P architecture; Baidu lit a 30,000-card P800 cluster in H1 2025 and ran DeepSeek V3/R1 671B on a single 8-card box. The only independently verifiable compute figure is ~345 TFLOPS FP16 (dense), from a TechInsights die floorplan analysis — stored here; all other precisions are not published and are omitted. Memory type is recorded as GDDR6 (consistent with the Kunlun II lineage and analyst teardowns); the "96GB HBM3" claims in marketing blogs are self-contradictory and rejected. Memory capacity/bandwidth, TDP, form factor and interconnect are not reliably disclosed and are left NULL.
Performance
Peak theoretical throughput by precision type
| Precision | Peak |
|---|---|
FP64 | No verified data available |
FP32 | No verified data available |
TF32 | No verified data available |
BF16 | No verified data available |
FP16 16-bit floating point | 345TFLOPS |
FP8 | No verified data available |
FP6 | No verified data available |
FP4 | No verified data available |
INT8 | No verified data available |
Specifications
Architecture
XPU-P (3rd-gen)
Form Factor
No verified data available
Launch Year
2024
Process Node
7nm
Memory
No verified data available GDDR6
Bandwidth
No verified data available
TDP
No verified data available
Max power
No verified data available
Spec Confidence
Vendor claimed
Full Specifications
| Memory | |
|---|---|
| Memory | No verified data available |
| Memory Type | GDDR6 |
| Bandwidth | No verified data available |
| Interface Width | No verified data available |
| Power & Thermal | |
| TDP | No verified data available |
| Max power | No verified data available |
| General | |
| Form Factor | No verified data available |
| Architecture | XPU-P (3rd-gen) |
| Process Node | 7nm |
| Launch Year | 2024 |
Datasheet & Resources
Flopper Datasheet
Baidu Kunlun P800 Kunlun III specifications, generated from our database. Printable.
Baidu Kunlunxin P800 AI Accelerator Processor Floorplan Analysis
TechInsights · Latest version
Data Provenance
Every figure traced to a source
Primary Source
- Publisher
- TechInsights
- Published
- No verified data available
Data Quality
- Spec confidence
- Vendor claimed
- Clock basis
- Boost
- Core precisions with figures
- 1 of 9
- Normalization
- All values in TFLOPS
Compare cloud providers offering on-demand GPU instances for AI training, inference, and HPC workloads.
Browse GPU Cloud ProvidersFrequently Asked Questions
How many TFLOPS does the Baidu Kunlun P800 have?
The Baidu Kunlun P800 delivers 345 TFLOPS FP16 at peak. Flopper does not currently have verified FP32 and FP8 throughput figures for it.
What is the power consumption of the Baidu Kunlun P800?
Flopper does not currently have a verified TDP (Thermal Design Power) figure for the Baidu Kunlun P800.
How much memory does the Baidu Kunlun P800 have?
Flopper does not currently have verified memory capacity or bandwidth figures for the Baidu Kunlun P800.
What architecture is the Baidu Kunlun P800 based on?
The Baidu Kunlun P800 is based on the XPU-P (3rd-gen) architecture, launched in 2024.
Stay Updated on GPU Releases
Get notified when new GPUs are added or specifications are updated.
No spam, unsubscribe anytime.