Back to Arena

Talkie AI

by MiniMax

System Card

OrganizationMiniMax
Released2023-06
Architectureepisodic-buffer / Long-Context Character Conversation
DetailsTalkie is built on MiniMax's M1 model with a massive context window enabling deep, character-faithful multi-turn conversations that remember details from early in a chat history. Character personas evolve and adapt with extended memory as users engage over time. The platform supports voice acting, scene-specific backgrounds, roleplay, and user character creation. Launched as an iteration of the Glow app in June 2023.
Parameters—
Domainpersonalizationepisodic-session
Open SourceNo
WebsiteVisit
companioncharacter-roleplayminimaxlong-contextvoice-actingimmersive

Capability Profile

Benchmark Scores

5 of 14 benchmarks
Long-Context Retrieval
0/5
RULER
no data
NIAH
no data
LooGLE
no data
LongBench
no data
∞Bench
no data
Multi-Turn Recall
2/2
LoCoMo
81.892p
MemoryBank
8593p
Cross-Session Memory
1/1
LongMemEval
82.891p
Multi-Hop QA
0/3
BABILong
no data
MultiHop-RAG
no data
HotpotQA
no data
Agent Task Memory
1/1
AgentBench-Mem
7226p
Personalization
1/1
PerLTQA
84.877p
Factuality / Grounding
0/1
RAGAS
no data
Sources:Talkie AI vendor documentation; evaluated on LoCoMo: Long-Term Conversational Memory Benchmark (Snap Research, 2402)Talkie AI vendor documentation; evaluated on LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory (Salesforce AI Research, 2410)Talkie AI vendor documentation; evaluated on MemoryBank: Enhancing LLMs with Long-Term Memory (Sun Yat-sen University, 2305)Talkie AI vendor documentation; evaluated on AgentBench Memory Track (Tsinghua KEG, 2308)Talkie AI vendor documentation; evaluated on PerLTQA: A Personal Long-Term Memory Question Answering Dataset (PolyU, 2402)