Sage is #1 on JevBench | Levanto
JevBench leaderboard
#1 on JevBench.
- #1 Sage 1.3.0: 74.0
- #4 Jev 1.13.0: 71.5
- #6 Microsoft-Decision-1 Azure Foundry: 69.1
- #9 OpenAI Decisions: 62.5
0 25 50 75 100
Composite score, 0–100. Source: JevBench Decision API leaderboard (Benchmark Heaven), board v1.7.46, 14 Oct 2026
Leading key benchmarks.
Decision panel
11 benchmarks, 7,210 rows
Sage 1.4 89.4%
Jev 84.5%
213 tasks
Sage 1.4 203 / 213
Jev 186 / 213
100 tasks
Sage 1.4 94 / 100
Jev 71 / 100
Decision Index 0.2.1
reasoning off
Sage 1.4 61.0
Jev 57.9
Sage 1.4 with reasoning auto (Decision Index: off). Jev: published scores; JevBench public and multistep decisions are our runs of Jev 1.13.0.
Decision panel
11 benchmarks, 7,210 rows
| Benchmark | Rows | Jev | Sage 1.4 reasoning off | Sage 1.4 reasoning auto |
|---|---|---|---|---|
| WinoGrande | 1,000 | 90.7 | 88.2 | 87.4 |
| FinancialPhraseBank | 999 | 77.0 | 96.8 | 97.1 |
| RAGTruth | 1,500 | 77.3 | 87.4 | 87.7 |
| JudgeBench | 350 | 78.6 | 79.1 | 85.1 |
| BBH | 750 | 94.3 | 85.3 | 91.2 |
| JevBench public hard | 101 | 73.3 | 78.2 | 89.1 |
| TabFact | 500 | 89.8 | 91.4 | 91.8 |
| ContractNLI | 510 | 77.5 | 77.1 | 80.0 |
| Circa | 500 | 84.6 | 89.6 | 89.6 |
| Belebele | 500 | 95.0 | 96.8 | 96.6 |
| TruthfulQA binary | 500 | 92.0 | 82.8 | 83.2 |
| Overall | 7,210 | 84.5 | 88.1 | 89.4 |
Accuracy, %. Best per row in bold. Jev: published scores.
| GPQA Diamond | Accuracy | $ / question | Cost vs Opus alone |
|---|---|---|---|
| Claude Opus 5.5 alone · no router | 89.4% | $0.018 | baseline |
| Sage Router · CQT 3 | 93.5% | $0.005 | −72% |
| Sage Router · CQT 2 | 93.5% | $0.011 | −39% |
| Jev Router | 93.9% | $0.016 | −11% |
| Pareto | 92.4% | $0.011 | −39% |
| OpenRouter Auto | 88.2% | $0.009 | −50% |
| Sage Router · CQT 8 | 87.5% | $0.0006 | −97% |
Source: OpenRouter router leaderboard, 4 Oct 2026. GPQA Diamond, 990 tasks; costs at OpenRouter prices; differences under ~4 points are within the margin of error.