YZ Index
Weekly Report
Weekly model performance changes and trend analysis.
Baseline: Run #419 · Formula v7 · Judge v6.5 · Benchmark v7 · 2026-11-09 05:06 SGT
Current: Run #430 · Formula v7 · Judge v6.5 · Benchmark v7 · 2026-11-16 05:16 SGT
Overall Score Changes Ranked by absolute change magnitude
Claude Sonnet 4.6
+4.3
71.3 → 75.6
GPT-6 Sol
+4.1
75.6 → 79.7
GPT-5.5
+3.9
73.3 → 77.2
Grok 4
+2.3
70.9 → 73.2
DeepSeek V4 Pro
+1.6
66.9 → 68.5
Doubao Pro
-8.0
80.5 → 72.5
Gemini 3.1 Pro
-5.4
62.9 → 57.5
GPT-6.1 Sol
-4.7
82.8 → 78.1
Gemini 2.5 Pro
-2.0
59.0 → 56.9
GPT-o3
-1.8
72.2 → 70.4
GPT-6 Luna
-0.9
76.9 → 76.0
2 — stable
Claude Opus 4.7 (79.5)
GPT-6 Astra (80.3)
Side Dimension Changes Communication and Judgment changes
DeepSeek V4 Pro
+31.3
Judgment: 56.5 → 87.8
DeepSeek V4 Pro
+20.4
Communication: 63.8 → 84.2
GPT-6 Sol
+11.9
Communication: 74.2 → 86.1
Claude Sonnet 4.6
+11.1
Judgment: 58.6 → 69.7
Gemini 2.5 Pro
+10.1
Judgment: 70.0 → 80.1
Claude Opus 4.7
+5.6
Judgment: 68.0 → 73.6
Gemini 3.1 Pro
+5.0
Judgment: 77.8 → 82.8
Claude Opus 4.7
+2.5
Communication: 63.3 → 65.8
GPT-6 Sol
+2.5
Judgment: 86.5 → 89.0
Grok 4
-8.3
Communication: 74.8 → 66.5
Doubao Pro
-6.2
Communication: 93.3 → 87.1
GPT-6 Luna
-5.2
Communication: 84.6 → 79.4
Claude Sonnet 4.6
-4.3
Communication: 68.1 → 63.8
GPT-o3
-3.3
Judgment: 86.7 → 83.4
Grok 4
-3.2
Judgment: 47.3 → 44.1
GPT-5.5
-3.1
Communication: 90.8 → 87.7
Gemini 3.1 Pro
-2.5
Communication: 78.3 → 75.8
GPT-6.1 Sol
-2.5
Communication: 88.8 → 86.3
GPT-o3
-2.5
Communication: 78.3 → 75.8
GPT-5.5
-2.2
Judgment: 90.0 → 87.8
Gemini 2.5 Pro
-2.0
Communication: 78.5 → 76.5
GPT-6 Luna
-2.0
Judgment: 88.5 → 86.5
Operational Signal Changes Stability and Availability changes
GLM-4.6
+80.7
Availability: 0.0 → 80.7
GLM-4.6
+41.4
Value: 0.0 → 41.4
GPT-6 Sol
+8.7
Stability: 35.1 → 43.8
DeepSeek V4 Pro
+8.0
Availability: 82.8 → 90.8
GPT-5.5
+7.5
Stability: 35.1 → 42.6
Claude Sonnet 4.6
+6.2
Stability: 25.5 → 31.7
Gemini 2.5 Pro
+5.1
Stability: 31.3 → 36.4
DeepSeek V4 Pro
+3.8
Stability: 40.7 → 44.5
DeepSeek V4 Pro
+3.8
Value: 42.0 → 45.8
Claude Opus 4.7
+2.4
Stability: 31.2 → 33.6
GPT-6 Sol
+1.6
Value: 37.2 → 38.8
GPT-5.5
+1.0
Availability: 97.0 → 98.0
Grok 4
+0.9
Stability: 27.4 → 28.3
Claude Sonnet 4.6
+0.8
Value: 23.5 → 24.3
GPT-5.5
+0.5
Value: 18.4 → 18.9
GLM-4.6
-33.9
Stability: 75.0 → 41.1
GPT-o3
-5.5
Stability: 42.8 → 37.3
GPT-6.1 Sol
-4.8
Stability: 42.9 → 38.1
Doubao Pro
-4.7
Stability: 43.0 → 38.3
Gemini 2.5 Pro
-3.9
Availability: 86.9 → 83.0
GPT-6 Astra
-2.0
Stability: 42.2 → 40.2
GPT-o3
-1.9
Availability: 92.9 → 91.0
GPT-6.1 Sol
-1.7
Value: 39.8 → 38.1
GPT-6 Luna
-1.6
Stability: 37.7 → 36.1
Doubao Pro
-1.2
Value: 95.8 → 94.6
Doubao Pro
-1.1
Availability: 98.9 → 97.8
Gemini 3.1 Pro
-1.0
Value: 22.5 → 21.5
GPT-6 Luna
-1.0
Availability: 100.0 → 99.0
Show legacy dimension changes
4
Up
7
Down
0
Stable
11
models
Significant Increases
Significant Decreases
GLM-4.6
-23.8
GLM-4.6: Engineering Judgment -23.8
judgment_raw
Gemini 2.5 Pro
-8.1
Gemini 2.5 Pro: Code Execution -8.1
execution_raw
Claude Sonnet 4.6
-7.9
Claude Sonnet 4.6: Code Execution -7.9
execution_raw
Claude Opus 4.7
-6.4
Claude Opus 4.7: Code Execution -6.4
execution_raw
GPT-o3
-4.4
GPT-o3: Code Execution -4.4
execution_raw
DeepSeek V4 Pro
-2.9
DeepSeek V4 Pro: Grounding -2.9
grounding_raw
GPT-5.5
-2.2
GPT-5.5: Code Execution -2.2
execution_raw