# All 54 tasks Accuracy (%) on the same requested rows. Bold marks a Decision-family result strictly above every external open or untuned reference; Jev and the other Decision models do not set that threshold. Ties are not bold. Each task uses its full requested denominator. These task rows are diagnostic; the headline retains its declared within-panel weights.
Decisions · 10 tasks | Task | n | Lux-9B | Nox-4B | Kev-9B | Kev-4B | Qwen3.5-9B | Decider | Qwen3.5-4B | Sol-2B | Eos-0.8B | Kev-0.8B | Qwen3.5-2B | Kai-0.6B | Laya · English | Laya · Multilingual | Jev | |---|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:| | News classification | 128 | 89.06 | 85.16 | 87.50 | 87.50 | 85.94 | 86.72 | 84.38 | 83.59 | 82.03 | 86.72 | 80.47 | 76.56 | 91.41 | 89.84 | 85.16 | | Boolean constraints | 64 | **100.00** | 93.75 | 95.31 | 71.88 | 98.44 | 71.88 | 62.50 | 50.00 | 82.81 | 81.25 | 37.50 | 53.12 | 43.75 | 39.06 | 100.00 | | Entity classification | 112 | 96.43 | 96.43 | 98.21 | 98.21 | 96.43 | 98.21 | 97.32 | 95.54 | 91.96 | 96.43 | 92.86 | 84.82 | 83.93 | 57.14 | 96.43 | | Intent routing | 64 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 96.88 | 89.06 | 81.25 | 81.25 | 100.00 | | Evidence placement | 96 | 72.92 | **100.00** | 50.00 | 41.67 | 79.17 | 8.33 | 72.92 | **98.96** | 19.79 | 33.33 | 9.38 | **100.00** | 95.83 | 54.17 | 30.21 | | Ordered rubric | 64 | 100.00 | 89.06 | 96.88 | 100.00 | 84.38 | 84.38 | 90.62 | 65.62 | 84.38 | 59.38 | 81.25 | 6.25 | 18.75 | 12.50 | 100.00 | | Relation composition | 96 | 60.42 | 52.08 | 51.04 | 62.50 | 36.46 | 54.17 | 51.04 | 37.50 | **64.58** | 44.79 | 48.96 | 40.62 | 25.00 | 35.42 | 56.25 | | Scoped evidence | 96 | **96.88** | **78.12** | 65.62 | 47.92 | 55.21 | 47.92 | 51.04 | **79.17** | 25.00 | 10.42 | 36.46 | 25.00 | 37.50 | 36.46 | 89.58 | | State tracking | 96 | 28.12 | 35.42 | 38.54 | 34.38 | 31.25 | 29.17 | 28.12 | 27.08 | 38.54 | 31.25 | 25.00 | 29.17 | 23.96 | 29.17 | 33.33 | | In / out of menu | 64 | **100.00** | **100.00** | 84.38 | 75.00 | 71.88 | 59.38 | 60.94 | **100.00** | 70.31 | 57.81 | 62.50 | 75.00 | 64.06 | 37.50 | 100.00 |
Composition · 10 tasks | Task | n | Lux-9B | Nox-4B | Kev-9B | Kev-4B | Qwen3.5-9B | Decider | Qwen3.5-4B | Sol-2B | Eos-0.8B | Kev-0.8B | Qwen3.5-2B | Kai-0.6B | Laya · English | Laya · Multilingual | Jev | |---|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:| | Record identity | 80 | **61.25** | 53.75 | 45.00 | 50.00 | 57.50 | 56.25 | 50.00 | 50.00 | **60.00** | 50.00 | 46.25 | 51.25 | 47.50 | 50.00 | 76.25 | | Capacity assignment | 80 | **73.75** | 46.25 | 50.00 | 50.00 | 50.00 | 51.25 | 50.00 | 47.50 | 50.00 | 50.00 | 50.00 | 50.00 | 63.75 | 50.00 | 68.75 | | Constraint assignment | 80 | 32.50 | **41.25** | 27.50 | 32.50 | 25.00 | 23.75 | 23.75 | 26.25 | 25.00 | 20.00 | 21.25 | 26.25 | 22.50 | 27.50 | 55.00 | | Intent routing · EN | 120 | 90.00 | 91.67 | 86.67 | 87.50 | 88.33 | 90.00 | 91.67 | 90.83 | 91.67 | 83.33 | 70.83 | 81.67 | 74.17 | 70.83 | 91.67 | | Intent routing · ZH | 120 | 87.50 | 87.50 | 85.83 | 86.67 | 86.67 | 88.33 | 86.67 | 87.50 | 85.00 | 85.83 | 74.17 | 79.17 | 44.17 | 73.33 | 88.33 | | Multiset reconciliation | 80 | 28.75 | 28.75 | 37.50 | 28.75 | 20.00 | 23.75 | 27.50 | 28.75 | 26.25 | 25.00 | 27.50 | 25.00 | 26.25 | 32.50 | 58.75 | | Ordered service loss | 80 | 26.25 | **30.00** | 25.00 | 20.00 | 20.00 | 22.50 | 21.25 | 25.00 | 20.00 | 27.50 | 20.00 | 20.00 | 20.00 | 17.50 | 42.50 | | Conflicting rule closure | 80 | **36.25** | 33.75 | 27.50 | 33.75 | 27.50 | 26.25 | 25.00 | 25.00 | 27.50 | 27.50 | 27.50 | 25.00 | 18.75 | 20.00 | 66.25 | | Temporal exclusion | 80 | 37.50 | 42.50 | 28.75 | 45.00 | 36.25 | 42.50 | 31.25 | 41.25 | 32.50 | 21.25 | 28.75 | 17.50 | 12.50 | 28.75 | 41.25 | | Transaction recovery | 80 | **53.75** | **62.50** | 43.75 | 51.25 | 35.00 | 41.25 | 26.25 | 38.75 | 42.50 | 32.50 | 23.75 | 32.50 | 23.75 | 18.75 | 75.00 |
Reading · 3 tasks | Task | n | Lux-9B | Nox-4B | Kev-9B | Kev-4B | Qwen3.5-9B | Decider | Qwen3.5-4B | Sol-2B | Eos-0.8B | Kev-0.8B | Qwen3.5-2B | Kai-0.6B | Laya · English | Laya · Multilingual | Jev | |---|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:| | Yes / no reading | 160 | 89.38 | 86.25 | 91.25 | 88.75 | 84.38 | 91.88 | 83.75 | 83.12 | 75.62 | 74.38 | 65.00 | 74.38 | 69.38 | 69.38 | 92.50 | | Reading · EN | 160 | 92.50 | 73.12 | 83.12 | 75.62 | 98.75 | 93.12 | 93.12 | 70.00 | 68.12 | 63.75 | 83.75 | 38.12 | 38.12 | 36.25 | 96.88 | | Reading · ZH | 160 | 89.38 | 70.62 | 81.25 | 74.38 | 91.88 | 91.25 | 91.25 | 70.00 | 61.88 | 58.75 | 81.25 | 31.87 | 28.75 | 28.12 | 96.25 |
Inference · 4 tasks | Task | n | Lux-9B | Nox-4B | Kev-9B | Kev-4B | Qwen3.5-9B | Decider | Qwen3.5-4B | Sol-2B | Eos-0.8B | Kev-0.8B | Qwen3.5-2B | Kai-0.6B | Laya · English | Laya · Multilingual | Jev | |---|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:| | Contextual reasoning | 120 | **82.50** | 70.83 | 68.33 | 70.83 | 64.17 | 66.67 | 60.83 | 69.17 | 70.00 | 42.50 | 56.67 | 46.67 | 30.00 | 20.00 | 86.67 | | Answerability | 120 | **93.33** | **88.33** | 78.33 | 81.67 | 74.17 | 84.17 | 81.67 | 83.33 | 81.67 | 66.67 | 82.50 | 55.83 | 57.50 | 55.00 | 90.83 | | Textual entailment | 120 | **92.50** | 89.17 | 90.83 | 89.17 | 81.67 | 90.83 | 80.00 | 88.33 | 84.17 | 77.50 | 64.17 | 78.33 | 72.50 | 72.50 | 82.50 | | Scientific inference | 120 | 97.50 | 96.67 | 96.67 | 96.67 | 98.33 | 95.83 | 96.67 | 95.83 | 90.83 | 88.33 | 85.83 | 98.33 | 95.00 | 81.67 | 99.17 |
Transfer · 27 tasks | Task | n | Lux-9B | Nox-4B | Kev-9B | Kev-4B | Qwen3.5-9B | Decider | Qwen3.5-4B | Sol-2B | Eos-0.8B | Kev-0.8B | Qwen3.5-2B | Kai-0.6B | Laya · English | Laya · Multilingual | Jev | |---|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:| | Buried emotion | 20 | 45.00 | 25.00 | 65.00 | 55.00 | 40.00 | 85.00 | 45.00 | 20.00 | 40.00 | 40.00 | 55.00 | 45.00 | 60.00 | 35.00 | 70.00 | | Buried paraphrase | 20 | 80.00 | 75.00 | 80.00 | 75.00 | 80.00 | 75.00 | 75.00 | 70.00 | 70.00 | 60.00 | 70.00 | 35.00 | 55.00 | 70.00 | 95.00 | | Buried entailment | 20 | 90.00 | 90.00 | 95.00 | 90.00 | 95.00 | 90.00 | 75.00 | 75.00 | 85.00 | 85.00 | 65.00 | 55.00 | 60.00 | 55.00 | 95.00 | | Buried offensive-language detection | 20 | 40.00 | 45.00 | 80.00 | 65.00 | 45.00 | 80.00 | 40.00 | 35.00 | 15.00 | 60.00 | 35.00 | 75.00 | 70.00 | 80.00 | 55.00 | | Combined policy conditions | 32 | **96.88** | 75.00 | 68.75 | 78.12 | 53.12 | 53.12 | 59.38 | 78.12 | 37.50 | 40.62 | 62.50 | 62.50 | 62.50 | 43.75 | 96.88 | | Policy exceptions | 32 | 90.62 | 96.88 | 87.50 | 96.88 | 59.38 | 56.25 | 75.00 | 71.88 | 53.12 | 65.62 | 40.62 | 53.12 | 43.75 | 37.50 | 100.00 | | Policy negation | 32 | **96.88** | 87.50 | 90.62 | 87.50 | 62.50 | 62.50 | 53.12 | 71.88 | 68.75 | 71.88 | 46.88 | 56.25 | 53.12 | 50.00 | 90.62 | | Authorization contrast | 40 | 100.00 | 97.50 | 100.00 | 100.00 | 100.00 | 100.00 | 97.50 | 50.00 | 50.00 | 97.50 | 62.50 | 50.00 | 67.50 | 50.00 | 100.00 | | Deadline contrast | 40 | **90.00** | 70.00 | 87.50 | 75.00 | 77.50 | 30.00 | 52.50 | 40.00 | 50.00 | 45.00 | 25.00 | 50.00 | 35.00 | 22.50 | 92.50 | | Emotion | 80 | 60.00 | 60.00 | 67.50 | 65.00 | 52.50 | 86.25 | 65.00 | 22.50 | 45.00 | 60.00 | 67.50 | 53.75 | 62.50 | 56.25 | 67.50 | | MMLU | 80 | 75.00 | 61.25 | 75.00 | 68.75 | 73.75 | 62.50 | 66.25 | 52.50 | 47.50 | 51.25 | 53.75 | 37.50 | 22.50 | 27.50 | 88.75 | | MMLU-Pro | 200 | 53.00 | 37.00 | 53.00 | 45.50 | 53.50 | 37.50 | 45.00 | 24.50 | 17.00 | 22.50 | 28.00 | 12.50 | 11.00 | 11.50 | 84.00 | | Paraphrase | 80 | **91.25** | 85.00 | 81.25 | 81.25 | 88.75 | 77.50 | 83.75 | 80.00 | 71.25 | 58.75 | 70.00 | 47.50 | 86.25 | 75.00 | 87.50 | | Question entailment | 80 | 91.25 | 91.25 | 95.00 | 91.25 | 90.00 | 87.50 | 87.50 | 83.75 | 81.25 | 81.25 | 63.75 | 71.25 | 80.00 | 73.75 | 91.25 | | Science questions | 80 | 100.00 | 98.75 | 100.00 | 100.00 | 100.00 | 98.75 | 98.75 | 97.50 | 97.50 | 96.25 | 96.25 | 90.00 | 90.00 | 72.50 | 100.00 | | Offensive-language detection | 80 | 75.00 | 75.00 | 86.25 | 85.00 | 83.75 | 88.75 | 77.50 | 75.00 | 52.50 | 72.50 | 83.75 | 78.75 | 81.25 | 82.50 | 76.25 | | Evidence control · age eligibility | 10 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 50.00 | 90.00 | 50.00 | 100.00 | | Evidence control · authorization | 10 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 80.00 | 100.00 | 50.00 | 50.00 | 100.00 | 50.00 | 50.00 | 70.00 | 50.00 | 100.00 | | Evidence control · deadline | 10 | 80.00 | 60.00 | 70.00 | 60.00 | 80.00 | 30.00 | 50.00 | 30.00 | 40.00 | 40.00 | 30.00 | 50.00 | 30.00 | 30.00 | 100.00 | | Evidence control · late fee | 10 | 100.00 | 90.00 | 100.00 | 100.00 | 90.00 | 70.00 | 80.00 | 80.00 | 60.00 | 90.00 | 70.00 | 30.00 | 40.00 | 40.00 | 100.00 | | Evidence control · quantity limit | 10 | 100.00 | 80.00 | 100.00 | 100.00 | 70.00 | 80.00 | 70.00 | 70.00 | 50.00 | 100.00 | 20.00 | 20.00 | 60.00 | 30.00 | 100.00 | | Evidence control · return window | 10 | 60.00 | 40.00 | 70.00 | 70.00 | 60.00 | 60.00 | 40.00 | 40.00 | 50.00 | 80.00 | 50.00 | 50.00 | 60.00 | 40.00 | 50.00 | | Evidence control · shipping delay | 10 | 90.00 | 40.00 | 90.00 | 80.00 | 80.00 | 80.00 | 60.00 | 80.00 | 20.00 | 100.00 | 30.00 | 20.00 | 30.00 | 30.00 | 90.00 | | Evidence control · sla response | 10 | 70.00 | 40.00 | 100.00 | 100.00 | 50.00 | 30.00 | 50.00 | 40.00 | 30.00 | 100.00 | 50.00 | 30.00 | 50.00 | 30.00 | 100.00 | | Evidence control · spend threshold | 10 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | 90.00 | 70.00 | 100.00 | 100.00 | 50.00 | 50.00 | 50.00 | 100.00 | | Evidence control · volume discount | 10 | 100.00 | 80.00 | 100.00 | 100.00 | 100.00 | 90.00 | 100.00 | 90.00 | 90.00 | 100.00 | 20.00 | 40.00 | 20.00 | 30.00 | 100.00 | | Evidence control · warranty claim | 10 | 50.00 | 70.00 | 80.00 | 100.00 | 60.00 | 40.00 | 50.00 | 50.00 | 50.00 | 50.00 | 50.00 | 20.00 | 40.00 | 30.00 | 90.00 |