Cohere · cohere/tiny-aya-global
| Provider | Cohere |
|---|---|
| Family | aya |
| Type | llm-chat |
| Status | active |
| Released | 2026-02-13 |
| Updated | 2026-07-07 |
| Parameters | 3.3B |
| Open weights | yes |
| Relationship | Model |
|---|---|
| Is a finetune of | CohereLabs/tiny-aya-base |
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 | 6.7 GB | ~2194535.9 | ~8778143.6 q4 |
| NVIDIA GB200 Grace Blackwell Superchip | 372.0 GB | 16000.0 GB/s | bf16 | 6.7 GB | ~1672.0 | ~6688.1 q4 |
| AMD Instinct MI355X | 288.0 GB | 8000.0 GB/s | bf16 | 6.7 GB | ~836.0 | ~3344.1 q4 |
| NVIDIA B200 | 180.0 GB | 7700.0 GB/s | bf16 | 6.7 GB | ~804.7 | ~3218.7 q4 |
| Google TPU7x (Ironwood) | 192.0 GB | 7380.0 GB/s | bf16 | 6.7 GB | ~771.2 | ~3084.9 q4 |
| AMD Instinct MI325X | 256.0 GB | 6000.0 GB/s | bf16 | 6.7 GB | ~627.0 | ~2508.0 q4 |
| AMD Instinct MI300X | 192.0 GB | 5300.0 GB/s | bf16 | 6.7 GB | ~553.9 | ~2215.4 q4 |
| NVIDIA H200 SXM | 141.0 GB | 4800.0 GB/s | bf16 | 6.7 GB | ~501.6 | ~2006.4 q4 |
| NVIDIA H100 NVL | 94.0 GB | 3900.0 GB/s | bf16 | 6.7 GB | ~407.6 | ~1630.2 q4 |
| NVIDIA H100 SXM | 80.0 GB | 3350.0 GB/s | bf16 | 6.7 GB | ~350.1 | ~1400.3 q4 |
| AMD Instinct MI250X | 128.0 GB | 3200.0 GB/s | bf16 | 6.7 GB | ~334.4 | ~1337.6 q4 |
| Google TPU v5p | 95.0 GB | 2765.0 GB/s | bf16 | 6.7 GB | ~288.9 | ~1155.8 q4 |
| NVIDIA A100 80GB SXM | 80.0 GB | 2039.0 GB/s | bf16 | 6.7 GB | ~213.1 | ~852.3 q4 |
| NVIDIA H100 PCIe | 80.0 GB | 2000.0 GB/s | bf16 | 6.7 GB | ~209.0 | ~836.0 q4 |
| NVIDIA GeForce RTX 5090 | 32.0 GB | 1792.0 GB/s | bf16 | 6.7 GB | ~187.3 | ~749.1 q4 |
| Google TPU v6e (Trillium) | 32.0 GB | 1638.0 GB/s | bf16 | 6.7 GB | ~171.2 | ~684.7 q4 |
| AMD Instinct MI210 | 64.0 GB | 1600.0 GB/s | bf16 | 6.7 GB | ~167.2 | ~668.8 q4 |
| NVIDIA A100 40GB SXM | 40.0 GB | 1555.0 GB/s | bf16 | 6.7 GB | ~162.5 | ~650.0 q4 |
| Google TPU v4 | 32.0 GB | 1200.0 GB/s | bf16 | 6.7 GB | ~125.4 | ~501.6 q4 |
| NVIDIA GeForce RTX 4090 | 24.0 GB | 1008.0 GB/s | bf16 | 6.7 GB | ~105.3 | ~421.4 q4 |
| AMD Radeon RX 7900 XTX | 24.0 GB | 960.0 GB/s | bf16 | 6.7 GB | ~100.3 | ~401.3 q4 |
| NVIDIA GeForce RTX 5080 | 16.0 GB | 960.0 GB/s | bf16 | 6.7 GB | ~100.3 | ~401.3 q4 |
| NVIDIA GeForce RTX 3090 | 24.0 GB | 936.0 GB/s | bf16 | 6.7 GB | ~97.8 | ~391.3 q4 |
| NVIDIA L40S | 48.0 GB | 864.0 GB/s | bf16 | 6.7 GB | ~90.3 | ~361.2 q4 |
| Apple M3 Ultra | 512.0 GB | 819.0 GB/s | bf16 | 6.7 GB | ~85.6 | ~342.3 q4 |
| AMD Radeon RX 7900 XT | 20.0 GB | 800.0 GB/s | bf16 | 6.7 GB | ~83.6 | ~334.4 q4 |
| Apple M2 Ultra | 192.0 GB | 800.0 GB/s | bf16 | 6.7 GB | ~83.6 | ~334.4 q4 |
| Google TPU v5e | 16.0 GB | 800.0 GB/s | bf16 | 6.7 GB | ~83.6 | ~334.4 q4 |
| NVIDIA GeForce RTX 4080 SUPER | 16.0 GB | 736.0 GB/s | bf16 | 6.7 GB | ~76.9 | ~307.7 q4 |
| NVIDIA GeForce RTX 4070 Ti SUPER | 16.0 GB | 672.0 GB/s | bf16 | 6.7 GB | ~70.2 | ~280.9 q4 |
| AMD Radeon RX 9070 XT | 16.0 GB | 640.0 GB/s | bf16 | 6.7 GB | ~66.9 | ~267.5 q4 |
| Apple M4 Max | 128.0 GB | 546.0 GB/s | bf16 | 6.7 GB | ~57.1 | ~228.2 q4 |
| Apple M1 Max | 64.0 GB | 400.0 GB/s | bf16 | 6.7 GB | ~41.8 | ~167.2 q4 |
| Apple M2 Max | 96.0 GB | 400.0 GB/s | bf16 | 6.7 GB | ~41.8 | ~167.2 q4 |
| Apple M3 Max | 128.0 GB | 400.0 GB/s | bf16 | 6.7 GB | ~41.8 | ~167.2 q4 |
| NVIDIA GeForce RTX 3060 12GB | 12.0 GB | 360.0 GB/s | bf16 | 6.7 GB | ~37.6 | ~150.5 q4 |
| NVIDIA L4 | 24.0 GB | 300.0 GB/s | bf16 | 6.7 GB | ~31.4 | ~125.4 q4 |
| NVIDIA GeForce RTX 4060 Ti 16GB | 16.0 GB | 288.0 GB/s | bf16 | 6.7 GB | ~30.1 | ~120.4 q4 |
| Apple M4 Pro | 64.0 GB | 273.0 GB/s | bf16 | 6.7 GB | ~28.5 | ~114.1 q4 |
| Apple M4 | 32.0 GB | 120.0 GB/s | bf16 | 6.7 GB | ~12.5 | ~50.2 q4 |
Taken from the card's availability section.
| Platform | Id on that platform | |
|---|---|---|
| Hugging Face | CohereLabs/tiny-aya-global | gated |
| Benchmark | Evaluated as | Score | Evidence date | Source kind | |
|---|---|---|---|---|---|
| aa_lcr | Tiny Aya Global | 0.0% | 2026-09-10 evaluated | independent evaluator | source |
| critpt | Tiny Aya Global | 0.0% | 2026-09-10 evaluated | independent evaluator | source |
| gpqa_diamond | Tiny Aya Global | 30.51% | 2026-09-10 evaluated | independent evaluator | source |
This card reports no benchmark scores yet.
| Benchmark | Catalogue standing | Score |
|---|