Instruction Following: IFLONG26 Leaderboard
| Models | Average | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ACHAP-CollarF1 | ACHAP-COMET | ACHAP-GC-BERTScore | ACHAP-TM-BERTScore | ACHAP-TM-MATCHED | ACHAP-WER | ASR-WER | QA-BERTScore | QE-accuracy | QE-format-accuracy | SUM-BERTScore | TRANS-COMET | |
| Qwen3-Omni | 0.304 | 0.283 | 0.708 | 0.741 | 0.381 | 0.552 | 0.540 | 0.294 | 0.454 | 0.500 | 0.103 | 0.446 |
| KIT_IWSLT26_IF_LONG_ unconstrained_ contrastive1 | 0.421 | 0.532 | 0.693 | 0.743 | 0.467 | 0.023 | 0.016 | 0.344 | 0.361 | 0.498 | 0.268 | 0.632 |
| Phi4-Multimodal | 0.104 | 0.248 | 0.666 | 0.749 | 0.224 | 0.985 | 0.820 | 0.402 | 0.394 | 0.253 | 0.170 | 0.411 |
| KIT_IWSLT26_IF_ LONG_ unconstrained_ primary | 0.474 | 0.566 | 0.737 | 0.772 | 0.525 | 0.078 | 0.067 | 0.427 | 0.250 | 0.001 | 0.275 | 0.564 |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_CONTRASTIVE_HYB | 0.000 | 0.516 | 0.611 | 0.799 | 0.132 | 0.050 | 0.032 | 0.368 | 0.290 | 0.351 | 0.202 | 0.526 |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_PRIMARY | 0.000 | 0.523 | 0.604 | 0.801 | 0.130 | 0.090 | 0.049 | 0.376 | 0.290 | 0.351 | 0.201 | 0.514 |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_CONTRASTIVE_VAD | 0.000 | 0.513 | 0.625 | 0.805 | 0.135 | 0.050 | 0.044 | 0.369 | 0.290 | 0.351 | 0.202 | 0.531 |
| Models | en → en | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ACHAP-CollarF1 | ACHAP-COMET | ACHAP-GC-BERTScore | ACHAP-TM-BERTScore | ACHAP-TM-MATCHED | ACHAP-WER | ASR-WER | QA-BERTScore | QE-accuracy | QE-format-accuracy | SUM-BERTScore | TRANS-COMET | |
| Qwen3-Omni | 0.609 | N/A | 0.877 | 0.882 | 0.607 | 0.208 | 0.158 | 0.327 | N/A | N/A | 0.187 | N/A |
| KIT_IWSLT26_IF_LONG_ unconstrained_ contrastive1 | 0.583 | N/A | 0.877 | 0.889 | 0.634 | 0.093 | 0.064 | 0.385 | N/A | N/A | 0.218 | N/A |
| Phi4-Multimodal | 0.084 | N/A | 0.837 | 0.844 | 0.247 | 0.941 | 0.281 | 0.420 | N/A | N/A | 0.183 | N/A |
| KIT_IWSLT26_IF_ LONG_ unconstrained_ primary | 0.436 | N/A | 0.869 | 0.882 | 0.481 | 0.311 | 0.269 | 0.412 | N/A | N/A | 0.212 | N/A |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_CONTRASTIVE_HYB | 0.000 | N/A | 0.803 | 0.884 | 0.135 | 0.200 | 0.126 | 0.377 | N/A | N/A | 0.156 | N/A |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_PRIMARY | 0.000 | N/A | 0.804 | 0.883 | 0.135 | 0.359 | 0.196 | 0.390 | N/A | N/A | 0.152 | N/A |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_CONTRASTIVE_VAD | 0.000 | N/A | 0.803 | 0.884 | 0.135 | 0.200 | 0.175 | 0.377 | N/A | N/A | 0.160 | N/A |
| Models | en → zh | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ACHAP-CollarF1 | ACHAP-COMET | ACHAP-GC-BERTScore | ACHAP-TM-BERTScore | ACHAP-TM-MATCHED | ACHAP-WER | ASR-WER | QA-BERTScore | QE-accuracy | QE-format-accuracy | SUM-BERTScore | TRANS-COMET | |
| Qwen3-Omni | 0.000 | 0.286 | 0.571 | 0.599 | 0.135 | N/A | N/A | 0.246 | 0.957 | 1.000 | 0.021 | 0.812 |
| KIT_IWSLT26_IF_LONG_ unconstrained_ contrastive1 | 0.103 | 0.451 | 0.511 | 0.556 | 0.221 | N/A | N/A | 0.360 | 0.739 | 0.993 | 0.378 | 0.847 |
| Phi4-Multimodal | 0.051 | 0.333 | 0.617 | 0.640 | 0.195 | N/A | N/A | 0.412 | 0.917 | 0.085 | 0.160 | 0.521 |
| KIT_IWSLT26_IF_ LONG_ unconstrained_ primary | 0.503 | 0.785 | 0.685 | 0.679 | 0.565 | N/A | N/A | 0.452 | 1.000 | 0.004 | 0.383 | 0.789 |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_CONTRASTIVE_HYB | 0.000 | 0.683 | 0.496 | 0.757 | 0.135 | N/A | N/A | 0.373 | 0.658 | 0.819 | 0.326 | 0.685 |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_PRIMARY | 0.000 | 0.681 | 0.496 | 0.757 | 0.135 | N/A | N/A | 0.380 | 0.658 | 0.819 | 0.319 | 0.655 |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_CONTRASTIVE_VAD | 0.000 | 0.698 | 0.496 | 0.757 | 0.135 | N/A | N/A | 0.374 | 0.658 | 0.819 | 0.325 | 0.699 |
| Models | en → it | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ACHAP-CollarF1 | ACHAP-COMET | ACHAP-GC-BERTScore | ACHAP-TM-BERTScore | ACHAP-TM-MATCHED | ACHAP-WER | ASR-WER | QA-BERTScore | QE-accuracy | QE-format-accuracy | SUM-BERTScore | TRANS-COMET | |
| Qwen3-Omni | 0.512 | 0.557 | 0.716 | 0.745 | 0.596 | N/A | N/A | 0.354 | N/A | N/A | 0.110 | 0.649 |
| KIT_IWSLT26_IF_LONG_ unconstrained_ contrastive1 | 0.489 | 0.842 | 0.709 | 0.768 | 0.499 | N/A | N/A | 0.322 | N/A | N/A | 0.269 | 0.841 |
| Phi4-Multimodal | 0.207 | 0.353 | 0.668 | 0.759 | 0.289 | N/A | N/A | 0.404 | N/A | N/A | 0.175 | 0.515 |
| KIT_IWSLT26_IF_ LONG_ unconstrained_ primary | 0.456 | 0.737 | 0.703 | 0.777 | 0.504 | N/A | N/A | 0.439 | N/A | N/A | 0.267 | 0.732 |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_CONTRASTIVE_HYB | 0.000 | 0.735 | 0.528 | 0.767 | 0.123 | N/A | N/A | 0.373 | N/A | N/A | 0.174 | 0.695 |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_PRIMARY | 0.000 | 0.723 | 0.532 | 0.773 | 0.123 | N/A | N/A | 0.385 | N/A | N/A | 0.185 | 0.707 |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_CONTRASTIVE_VAD | 0.000 | 0.719 | 0.586 | 0.790 | 0.135 | N/A | N/A | 0.376 | N/A | N/A | 0.171 | 0.702 |
| Models | en → de | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ACHAP-CollarF1 | ACHAP-COMET | ACHAP-GC-BERTScore | ACHAP-TM-BERTScore | ACHAP-TM-MATCHED | ACHAP-WER | ASR-WER | QA-BERTScore | QE-accuracy | QE-format-accuracy | SUM-BERTScore | TRANS-COMET | |
| Qwen3-Omni | 0.094 | 0.288 | 0.667 | 0.736 | 0.185 | N/A | N/A | 0.247 | 0.858 | 1.000 | 0.092 | 0.323 |
| KIT_IWSLT26_IF_LONG_ unconstrained_ contrastive1 | 0.508 | 0.836 | 0.676 | 0.761 | 0.514 | N/A | N/A | 0.308 | 0.705 | 1.000 | 0.208 | 0.840 |
| Phi4-Multimodal | 0.074 | 0.308 | 0.543 | 0.753 | 0.165 | N/A | N/A | 0.370 | 0.658 | 0.927 | 0.162 | 0.607 |
| KIT_IWSLT26_IF_ LONG_ unconstrained_ primary | 0.500 | 0.740 | 0.690 | 0.748 | 0.551 | N/A | N/A | 0.405 | 0.000 | 0.000 | 0.238 | 0.733 |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_CONTRASTIVE_HYB | 0.000 | 0.647 | 0.616 | 0.789 | 0.135 | N/A | N/A | 0.350 | 0.501 | 0.584 | 0.153 | 0.723 |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_PRIMARY | 0.000 | 0.690 | 0.584 | 0.793 | 0.127 | N/A | N/A | 0.348 | 0.501 | 0.584 | 0.149 | 0.694 |
| FBK_IWSLT26_IF_LONG_ CONSTRAINED_CONTRASTIVE_VAD | 0.000 | 0.635 | 0.616 | 0.789 | 0.135 | N/A | N/A | 0.351 | 0.501 | 0.584 | 0.153 | 0.722 |