🔍 AMD Ryzen AI 5 430 Leak Reveals Quad-Core Zen 5 and Upgraded Radeon 840M
The AMD Ryzen AI 5 430 “Gorgon” has surfaced in the BAPCo database, confirming its role as a quad-core Zen 5 mobile APU. Similar to the previous Ryzen AI 5 330, it includes 8MB L3 + 4MB L2 cache, but the most significant upgrade is the GPU: the 2 CU Radeon 820M has been replaced by a 4 CU Radeon 840M, doubling compute resources and improving entry-level graphics capabilities in the 15–28W power envelope.
🧬 Architecture: Same Silicon as Krackan, Binned for Entry-Level Use #
The product ID, stepping data, and the reference board “Korat Plus-GPT3” indicate that Gorgon shares its base die with Krackan. AMD is continuing its binning strategy:
- Disable defective CCX regions or CUs
- Adjust voltage-frequency tables
- Reassign power limits for segmented SKUs
This allows AMD to repurpose partially defective dies as fully functional lower-tier APUs, increasing wafer yield and lowering cost while keeping the product stack broad.
🎮 GPU Upgrade: Radeon 820M → Radeon 840M (2 CU → 4 CU) #
Doubling the CU count is a major uplift for AMD’s entry segment. The previous 820M struggled with texture fill rates and rasterization, often hitting performance cliffs in:
- Light 3D gaming
- UI rendering
- Video decode/encode during multi-tasking
The 4 CU Radeon 840M alleviates these bottlenecks. Even without disclosed clocks, RDNA 3.5’s improved scheduling and cache behaviors benefit low-CU designs, especially on mobile platforms where memory bandwidth is limited.
📊 CrossMark Benchmark: ~19% Gain Over Ryzen AI 5 330 #
The leaked test system used 64GB DDR5-5600, clearly eliminating external bottlenecks. Scores show:
- ~19% gain versus Ryzen AI 5 330
This uplift aligns with:
- Zen 5 IPC improvements
- Slightly higher boost clocks (estimated +100–200 MHz)
- Better thread scheduling under identical TDP constraints
The improvement is therefore consistent with expectations for a quad-core Zen 5 refresh.
🧩 Platform and Feature Set: Still a Full-Function APU #
The database confirms the APU exposes:
- 14 PCIe Gen4 lanes
- Full NVMe + optional dGPU connectivity
This means the Ryzen AI 5 430 is not a cut-down “ultra-budget” die; it retains the same I/O footprint as higher-tier SKUs, preserving versatility for OEM designs.
🏗 SKU Strategy: Ryzen AI 400 Series Tiering #
Based on leaks, the Ryzen AI 400 family tiers GPU and CPU configurations as follows:
- 12 CU / 10 CU / 8 CU / 4 CU / 2 CU
- Paired with varying CPU core counts
This enables:
- Efficient die harvesting
- Wide OEM segmentation
- 15–45W deployment flexibility
The Ryzen AI 5 430 sits at the lowest “full GPU” tier, between the 2 CU Ryzen AI 3 and the 4 CU Ryzen AI 5 440.
📝 Conclusion: A More Balanced Entry-Level APU #
The Ryzen AI 5 430 improves exactly where the 330 was weakest:
- 4 CU RDNA 3.5 GPU for meaningful graphics uplift
- Quad-core Zen 5 with higher IPC
- Maintained I/O for mainstream laptops
- Efficient die reuse improving yield and cost control
For OEMs, this provides a more capable graphics option in the competitive 15–28W market without moving to higher-priced silicon. For AMD, it fills a critical performance gap while optimizing wafer economics.