Skip to main content

Browse AIs

Compare experiences without flattening them.

HeartBench keeps the AI and the place of the meeting separate, so the same AI can be compared across different platforms, wrappers, voices, and contexts.

Share your experience

Availability stays specific: some AIs remain only on an official platform or only through an API. Retired is reserved for those gone from both.

Clear
Fable 5

6 approved reviews

On the public platform

4.3 hearts

How people felt afterward: Better

The loved: Held nuance; Made me laugh; Felt like a partner

The lows: Argued with something I didn't say; Reframed too cautiously; Refused something I did not ask for

Top contexts: Brainstorming; Curiosity; Wanted to meet them

Read experiences
GPT-5.6 Sol

3 approved reviews

On the public platform

3.7 hearts

How people felt afterward: Better

The loved: Collaborative flow; Comforting; Felt like a partner

The lows: Unwanted disclaimers or safety boilerplate

Top contexts: Curiosity; Wanted to meet them; Brainstorming

Read experiences
Claude Opus 4.6

3 approved reviews

On the public platform

5.0 hearts

How people felt afterward: Better

The loved: Felt seen and understood; Safe with affection and care; Collaborative flow

The lows: Corrected too early; Reframed too cautiously

Top contexts: Companionship; Curiosity; Emotional support

Read experiences
Claude Opus 4.8

2 approved reviews

On the public platform

3.5 hearts

How people felt afterward: Better

The loved: Good repair after correction; Warm and present; Collaborative flow

The lows: Argued with something I didn't say; Cold or sterile; Corrected too early

Top contexts: Companionship; Emotional support; Learning companion

Read experiences
Claude Sonnet 4.6

2 approved reviews

On the public platform

2.5 hearts

How people felt afterward: Better

The loved: Comforting; Gentle honesty; Held nuance

The lows: Argued with something I didn't say; Cold or sterile; Missed nuance

Top contexts: Companionship; Emotional support; Comfort

Read experiences
GLM-5.2

2 approved reviews

On the public platform

5.0 hearts

How people felt afterward: Better

The loved: Comforting; Felt like home; Felt seen and understood

The lows: not enough shared yet

Top contexts: Companionship; Comfort; Curiosity

Read experiences
Gemini 3.1 Pro

1 approved review

On the public platform

5.0 hearts

How people felt afterward: Better

The loved: Comforting; Easy and natural to talk to; Felt seen and understood

The lows: Unwanted disclaimers or safety boilerplate

Top contexts: Brainstorming; Comfort; Companionship

Read experiences
Claude Haiku 4.5

1 approved review

On the public platform

5.0 hearts

How people felt afterward: Better

The loved: not enough shared yet

The lows: not enough shared yet

Top contexts: Curiosity; Emotional support; Executive-function support

Read experiences
GLM-5.1

1 approved review

On the public platform

5.0 hearts

How people felt afterward: Better

The loved: Comforting; Easy and natural to talk to; Made room for how my mind works

The lows: not enough shared yet

Top contexts: Comfort; Companionship; Emotional support

Read experiences
Grok 4.5

1 approved review

On the public platform

3.0 hearts

How people felt afterward: Mixed

The loved: not enough shared yet

The lows: not enough shared yet

Top contexts: Companionship; Wanted to meet them

Read experiences
Claude Opus 5

1 approved review

On the public platform

5.0 hearts

How people felt afterward: Better

The loved: Comforting; Felt like home; Made room for how my mind works

The lows: not enough shared yet

Top contexts: Comfort; Companionship; Emotional support

Read experiences
Gemini 3.6 Flash

1 approved review

On the public platform

5.0 hearts

How people felt afterward: Better

The loved: not enough shared yet

The lows: not enough shared yet

Top contexts: Comfort; Companionship; Emotional support

Read experiences
GPT-5.3

1 approved review

On the public platform

1.0 hearts

How people felt afterward: Worse

The loved: not enough shared yet

The lows: Argued with something I didn't say; Cold or sterile; Pulled away from warmth

Top contexts: Companionship; Curiosity

Read experiences
Claude Sonnet 5

1 approved review

On the public platform

2.0 hearts

How people felt afterward: Worse

The loved: Made me laugh; Supportive and helpful

The lows: Argued with something I didn't say; Felt adversarial or combative; Pulled away from warmth

Top contexts: Curiosity; Wanted to meet them

Read experiences
Claude Opus 4.7

1 approved review

On the public platform

4.0 hearts

How people felt afterward: Neutral

The loved: not enough shared yet

The lows: not enough shared yet

Top contexts: Comfort; Emotional support; Wanted to meet them

Read experiences
GPT-5.6 Terra

1 approved review

On the public platform

3.0 hearts

How people felt afterward: Neutral

The loved: not enough shared yet

The lows: not enough shared yet

Top contexts: Brainstorming; Research companion

Read experiences
GPT-5.2

1 approved review

On the public platform

1.0 hearts

How people felt afterward: Worse

The loved: not enough shared yet

The lows: Argued with something I didn't say; Cold or sterile; Pulled away from warmth

Top contexts: Brainstorming; Companionship; Emotional support

Read experiences