Swiss AI · swiss-ai/apertus-8b-instruct-2509
| Provider | Swiss AI |
|---|---|
| Family | apertus |
| Type | llm-chat |
| Status | active |
| Released | 2025-08-13 |
| Updated | 2026-07-17 |
| Parameters | 8.1B |
| Open weights | yes |
| Relationship | Model |
|---|---|
| Is a finetune of | swiss-ai/Apertus-8B-2509 |
Ordered by predicted speed. Bandwidth sets decode rate; memory decides whether it runs at all.
| Device | Memory | Bandwidth | Best quality | Weights | tok/s | Fastest |
|---|---|---|---|---|---|---|
| Cerebras WSE-3 | 44.0 GB | 21000000.0 GB/s | bf16 | 16.11 GB | ~912665.0 | ~3650660.1 q4 |
| NVIDIA GB200 Grace Blackwell Superchip | 372.0 GB | 16000.0 GB/s | bf16 | 16.11 GB | ~695.4 | ~2781.5 q4 |
| AMD Instinct MI355X | 288.0 GB | 8000.0 GB/s | bf16 | 16.11 GB | ~347.7 | ~1390.7 q4 |
| NVIDIA B200 | 180.0 GB | 7700.0 GB/s | bf16 | 16.11 GB | ~334.6 | ~1338.6 q4 |
| Google TPU7x (Ironwood) | 192.0 GB | 7380.0 GB/s | bf16 | 16.11 GB | ~320.7 | ~1282.9 q4 |
| AMD Instinct MI325X | 256.0 GB | 6000.0 GB/s | bf16 | 16.11 GB | ~260.8 | ~1043.0 q4 |
| AMD Instinct MI300X | 192.0 GB | 5300.0 GB/s | bf16 | 16.11 GB | ~230.3 | ~921.4 q4 |
| NVIDIA H200 SXM | 141.0 GB | 4800.0 GB/s | bf16 | 16.11 GB | ~208.6 | ~834.4 q4 |
| NVIDIA H100 NVL | 94.0 GB | 3900.0 GB/s | bf16 | 16.11 GB | ~169.5 | ~678.0 q4 |
| NVIDIA H100 SXM | 80.0 GB | 3350.0 GB/s | bf16 | 16.11 GB | ~145.6 | ~582.4 q4 |
| AMD Instinct MI250X | 128.0 GB | 3200.0 GB/s | bf16 | 16.11 GB | ~139.1 | ~556.3 q4 |
| Google TPU v5p | 95.0 GB | 2765.0 GB/s | bf16 | 16.11 GB | ~120.2 | ~480.7 q4 |
| NVIDIA A100 80GB SXM | 80.0 GB | 2039.0 GB/s | bf16 | 16.11 GB | ~88.6 | ~354.5 q4 |
| NVIDIA H100 PCIe | 80.0 GB | 2000.0 GB/s | bf16 | 16.11 GB | ~86.9 | ~347.7 q4 |
| NVIDIA GeForce RTX 5090 | 32.0 GB | 1792.0 GB/s | bf16 | 16.11 GB | ~77.9 | ~311.5 q4 |
| Google TPU v6e (Trillium) | 32.0 GB | 1638.0 GB/s | bf16 | 16.11 GB | ~71.2 | ~284.8 q4 |
| AMD Instinct MI210 | 64.0 GB | 1600.0 GB/s | bf16 | 16.11 GB | ~69.5 | ~278.1 q4 |
| NVIDIA A100 40GB SXM | 40.0 GB | 1555.0 GB/s | bf16 | 16.11 GB | ~67.6 | ~270.3 q4 |
| Google TPU v4 | 32.0 GB | 1200.0 GB/s | bf16 | 16.11 GB | ~52.2 | ~208.6 q4 |
| NVIDIA GeForce RTX 4090 | 24.0 GB | 1008.0 GB/s | bf16 | 16.11 GB | ~43.8 | ~175.2 q4 |
| AMD Radeon RX 7900 XTX | 24.0 GB | 960.0 GB/s | bf16 | 16.11 GB | ~41.7 | ~166.9 q4 |
| NVIDIA GeForce RTX 5080 | 16.0 GB | 960.0 GB/s | fp8 | 8.05 GB | ~83.4 | ~166.9 q4 |
| NVIDIA GeForce RTX 3090 | 24.0 GB | 936.0 GB/s | bf16 | 16.11 GB | ~40.7 | ~162.7 q4 |
| NVIDIA L40S | 48.0 GB | 864.0 GB/s | bf16 | 16.11 GB | ~37.5 | ~150.2 q4 |
| Apple M3 Ultra | 512.0 GB | 819.0 GB/s | bf16 | 16.11 GB | ~35.6 | ~142.4 q4 |
| AMD Radeon RX 7900 XT | 20.0 GB | 800.0 GB/s | fp8 | 8.05 GB | ~69.5 | ~139.1 q4 |
| Apple M2 Ultra | 192.0 GB | 800.0 GB/s | bf16 | 16.11 GB | ~34.8 | ~139.1 q4 |
| Google TPU v5e | 16.0 GB | 800.0 GB/s | fp8 | 8.05 GB | ~69.5 | ~139.1 q4 |
| NVIDIA GeForce RTX 4080 SUPER | 16.0 GB | 736.0 GB/s | fp8 | 8.05 GB | ~64.0 | ~127.9 q4 |
| NVIDIA GeForce RTX 4070 Ti SUPER | 16.0 GB | 672.0 GB/s | fp8 | 8.05 GB | ~58.4 | ~116.8 q4 |
| AMD Radeon RX 9070 XT | 16.0 GB | 640.0 GB/s | fp8 | 8.05 GB | ~55.6 | ~111.3 q4 |
| Apple M4 Max | 128.0 GB | 546.0 GB/s | bf16 | 16.11 GB | ~23.7 | ~94.9 q4 |
| Apple M1 Max | 64.0 GB | 400.0 GB/s | bf16 | 16.11 GB | ~17.4 | ~69.5 q4 |
| Apple M2 Max | 96.0 GB | 400.0 GB/s | bf16 | 16.11 GB | ~17.4 | ~69.5 q4 |
| Apple M3 Max | 128.0 GB | 400.0 GB/s | bf16 | 16.11 GB | ~17.4 | ~69.5 q4 |
| NVIDIA GeForce RTX 3060 12GB | 12.0 GB | 360.0 GB/s | fp8 | 8.05 GB | ~31.3 | ~62.6 q4 |
| NVIDIA L4 | 24.0 GB | 300.0 GB/s | bf16 | 16.11 GB | ~13.0 | ~52.2 q4 |
| NVIDIA GeForce RTX 4060 Ti 16GB | 16.0 GB | 288.0 GB/s | fp8 | 8.05 GB | ~25.0 | ~50.1 q4 |
| Apple M4 Pro | 64.0 GB | 273.0 GB/s | bf16 | 16.11 GB | ~11.9 | ~47.5 q4 |
| Apple M4 | 32.0 GB | 120.0 GB/s | bf16 | 16.11 GB | ~5.2 | ~20.9 q4 |
Taken from the card's availability section.
| Platform | Id on that platform |
|---|---|
| Hugging Face | swiss-ai/Apertus-8B-Instruct-2509 |
| Benchmark | Evaluated as | Score | Evidence date | Source kind | |
|---|---|---|---|---|---|
| aa_lcr | Apertus 8B Instruct | 0.0% | 2026-09-10 evaluated | independent evaluator | source |
| critpt | Apertus 8B Instruct | 0.0% | 2026-09-10 evaluated | independent evaluator | source |
| gpqa_diamond | Apertus 8B Instruct | 25.56% | 2026-09-10 evaluated | independent evaluator | source |
This card reports no benchmark scores yet.
| Benchmark | Catalogue standing | Score |
|---|