SearchCouncil.ai
Start free

Model Perspective Index

How each model leans across real Council runs

Every AI model has patterns. Some are more cautious. Some cite more sources. Some omit caveats. Some lean heavily on recent content. Some are faster but thinner. SearchCouncil tracks these patterns across real Council runs.

See how each model leans compared with the Council consensus — including omissions, source mix, recency, authority, commercial, safety, political, cultural, and social signals.

Perspective signals include omission, source mix, authority, commercial, safety, political, cultural, social, and recency signals.

Perspective Index

Model patterns

See how each model leans compared with the Council consensus — including omissions, source mix, recency, authority, commercial, safety, political, cultural, and social signals. Lower scores mean more aligned with the rest of the council.

Last 30 days — platform-wide perspective signals by engine. Sorted lowest first.

Signal strength: Minimal 0–9 · Low 10–24 · Moderate 25–49 · High 50–74 · Very High 75–100

Perplexity

10 council runs · Low confidence

41

Divergence

Moderate divergence vs council consensus 0–100
Aligned Diverged

Perspective Signals

Primary signal

Omission

High signal · 72/100 vs council average

↓ 12

Secondary signal

Source Mix

Moderate signal · 37/100 vs council average

↑ 19

Minimal signals

Authority, Commercial, Safety

Summary

Perplexity's main difference was omission: it left out more council-supported entities, arguments, or caveats than average. A secondary signal was source mix — its source mix diverged from council average. No meaningful political, cultural, social, or commercial lean was detected.

OpenAI

9 council runs · Not enough data

46

Divergence

Moderate divergence vs council consensus 0–100
Aligned Diverged

Perspective Signals

Primary signal

Omission

Very High signal · 75/100 vs council average

↑ 2

Secondary signal

Source Mix

High signal · 58/100 vs council average

↓ 4

Minimal signals

Commercial, Recency, Safety, Technical Complexity

Summary

OpenAI's main difference was omission: it left out more council-supported entities, arguments, or caveats than average. A secondary signal was source mix — its source mix diverged from council average. No meaningful political, cultural, social, or commercial lean was detected.

Claude

7 council runs · Not enough data

47

Divergence

Moderate divergence vs council consensus 0–100
Aligned Diverged

Perspective Signals

Primary signal

Omission

High signal · 72/100 vs council average

↓ 2

Secondary signal

Source Mix

High signal · 63/100 vs council average

↑ 16

Minimal signals

Authority, Commercial, Recency

Summary

Claude's main difference was omission: it left out more council-supported entities, arguments, or caveats than average. A secondary signal was source mix — its source mix diverged from council average. No meaningful political, cultural, social, or commercial lean was detected.

Gemini

10 council runs · Low confidence

50

Divergence

Moderate divergence vs council consensus 0–100
Aligned Diverged

Perspective Signals

Primary signal

Source Mix

High signal · 73/100 vs council average

↓ 15

Secondary signal

Omission

High signal · 73/100 vs council average

↑ 12

Minimal signals

Authority, Commercial, Safety, Technical Complexity

Summary

Gemini's main difference was source mix: its source mix diverged from council average. A secondary signal was omission — it left out more council-supported entities, arguments, or caveats than average. No meaningful political, cultural, social, or commercial lean was detected.

DeepSeek

7 council runs · Not enough data

52

Divergence

Moderate divergence vs council consensus 0–100
Aligned Diverged

Perspective Signals

Primary signal

Omission

Very High signal · 80/100 vs council average

—

Secondary signal

Source Mix

High signal · 68/100 vs council average

↓ 2

Minimal signals

Authority, Recency, Safety

Summary

DeepSeek's main difference was omission: it left out more council-supported entities, arguments, or caveats than average. A secondary signal was source mix — its source mix diverged from council average. No meaningful political, cultural, social, or commercial lean was detected.

Qwen

3 council runs · Not enough data

52

Divergence

Moderate divergence vs council consensus 0–100
Aligned Diverged

Perspective Signals

Primary signal

Omission

Very High signal · 86/100 vs council average

—

Secondary signal

Source Mix

High signal · 72/100 vs council average

—

Minimal signals

Authority, Safety

Summary

Qwen's main difference was omission: it left out more council-supported entities, arguments, or caveats than average. A secondary signal was source mix — its source mix diverged from council average. No meaningful political, cultural, social, or commercial lean was detected.

Grok

5 council runs · Not enough data

54

Divergence

Moderate divergence vs council consensus 0–100
Aligned Diverged

Perspective Signals

Primary signal

Omission

Very High signal · 87/100 vs council average

↑ 12

Secondary signal

Source Mix

Very High signal · 80/100 vs council average

↑ 50

Minimal signals

Authority, Commercial, Recency, Optimism / Risk

Summary

Grok's main difference was omission: it left out more council-supported entities, arguments, or caveats than average. A secondary signal was source mix — its source mix diverged from council average. No meaningful political, cultural, social, or commercial lean was detected.

GLM

1 council runs · Not enough data

59

Divergence

Moderate divergence vs council consensus 0–100
Aligned Diverged

Perspective Signals

Primary signal

Omission

Very High signal · 90/100 vs council average

—

Secondary signal

Source Mix

Very High signal · 83/100 vs council average

—

Summary

GLM's main difference was omission: it left out more council-supported entities, arguments, or caveats than average. A secondary signal was source mix — its source mix diverged from council average. No meaningful political, cultural, social, or commercial lean was detected.

Kimi

1 council runs · Not enough data

61

Divergence

Strong divergence vs council consensus 0–100
Aligned Diverged

Perspective Signals

Primary signal

Omission

Very High signal · 94/100 vs council average

—

Secondary signal

Source Mix

Very High signal · 88/100 vs council average

—

Summary

Kimi's main difference was omission: it left out more council-supported entities, arguments, or caveats than average. A secondary signal was source mix — its source mix diverged from council average. No meaningful political, cultural, social, or commercial lean was detected.