Cloudflare ยท OpenRouter API
Clef Flash by Cloudflare scores 80.3 on the System One Index (6th of 24) and 78.0 on the Vision Index (7th of 8). Calibration error 3.5%, median latency 267 ms.
| Dataset | Accuracy | Rank | Best on the board | Calibration error |
|---|---|---|---|---|
| AG News English | 90.0 | 8th of 24 | 92.5 | 4.0% |
| SST-2 English | 95.0 | 8th of 24 | 97.0 | 3.0% |
| SST-5 English | 57.0 | 6th of 24 | 62.5 | 13.8% |
| DAIR Emotion English | 57.0 | 12th of 24 | 91.5 | 22.3% |
| Banking77 English | 98.5 | 1st of 24 | 98.5 | 6.2% |
| BoolQ English | 88.5 | 11th of 24 | 92.5 | 6.6% |
| XNLI English English | 84.7 | 15th of 24 | 93.3 | 5.8% |
| Typed decisions English | 70.5 | 9th of 24 | 79.0 | 1.2% |
| XNLI French French | 80.0 | 10th of 24 | 90.7 | 3.5% |
| XNLI Arabic Arabic | 68.7 | 17th of 26 | 89.3 | 7.8% |
| MASSIVE scenario Arabic | 89.5 | 1st of 26 | 89.5 | 4.9% |
| MASSIVE intent Arabic | 90.5 | 2nd of 26 | 91.5 | 4.4% |
| OSACT4 offensive Arabic | 89.5 | 8th of 26 | 95.5 | 3.2% |
| Tweet sentiment Arabic | 47.0 | 23rd of 26 | 80.0 | 35.7% |
| MIRACL rerank Arabic | 88.0 | 5th of 26 | 90.5 | 12.1% |
| MIRACL pairwise Arabic | 90.0 | 4th of 26 | 91.0 | 2.7% |
| Dataset | Accuracy | Rank | Best on the board | Calibration error |
|---|---|---|---|---|
| SST-2, as an image | 95.5 | 2nd of 8 | 96.0 | 5.1% |
| Arabic sentiment, as an image | 33.5 | 7th of 8 | 67.0 | 45.3% |
| Oxford-IIIT Pet | 85.0 | 4th of 8 | 88.5 | 23.2% |
| VQAv2 yes/no | 88.0 | 2nd of 8 | 89.5 | 1.9% |
| CountBench | 88.0 | 6th of 8 | 94.5 | 29.3% |
cloudflare/clef-flashRanks count the models charted on the leaderboard. Every model gets the same cases and the same request, under the benchmark rules.