Early access: features, availability, and pricing may change as Kyara Intelligence evolves.
Compare models by capability, context window, and relative cost, automatically split into options that sit below and above the catalog baseline.
Lighter than the median model, more requests per unit of balance.
z-ai/glm-5.3-flash
Z.AI's fast multimodal reasoner for responsive, million-context RP.
deepseek/deepseek-v4-flash
DeepSeek's latest-generation model, setting a new bar for speed and intelligence in RP.
openai/gpt-6-luna
OpenAI's fast, cost-efficient GPT-6 reasoner with a 1M-token context window.
xiaomi/mimo-v2.5
Xiaomi's 1M-context omnimodal model for coherent, affordable long-form RP.
google/gemma-4-31b-it
Google DeepMind's dense multimodal Gemma model for coding, reasoning, and document understanding.
xiaomi/mimo-v2.6-flash
Xiaomi's efficient multimodal reasoner for affordable, long-context RP.
mistralai/mistral-small-2603
A signature model tuned for immersive RP and long-form storytelling.
stepfun/step-3.7-flash
A fast multimodal MoE with configurable reasoning for responsive, long-context RP.
openai/gpt-5.6-luna
OpenAI's fast, budget-friendly GPT-5.6 reasoner with a 1M-token context window.
arcee-ai/trinity-large-thinking
Large-scale open-weight MoE for creative writing and immersive roleplay adventures.
minimax/minimax-m3
MiniMax's multimodal million-context foundation model for ambitious long-form RP.
deepseek/deepseek-v4.1-flash
DeepSeek's first encoder-decoder Flash model with native vision and a 1M context.
meituan/longcat-2.0
Meituan's 1.6T sparse MoE for long-horizon reasoning and repository-scale creativity.
xiaomi/mimo-v2.5-pro
Xiaomi's flagship 1M-context model for coherent, ambitious, long-horizon RP.
xiaomi/mimo-v2.6-pro
Xiaomi's flagship multimodal model for ambitious, long-form RP.
deepseek/deepseek-v4-pro
DeepSeek's 1.6T flagship with advanced reasoning for deeply layered, long-running narratives.
deepseek/deepseek-v4-pro-0813
The GA release of DeepSeek's V4 Pro flagship, tuned for deep, long-running narratives.
Heavier than the median model, higher capability and more balance per call.
google/gemini-3.7-flash
Google's fast, SFW-friendly multimodal thinker with million-token memory for dynamic RP.
z-ai/glm-4.7
Z.AI's flagship with enhanced reasoning and natural conversation for sophisticated RP scenarios.
z-ai/glm-5
Z.AI's flagship open-source powerhouse for agentic reasoning and immersive, autonomous RP.
aion-labs/aion-3.0-mini
AionLabs' efficient multi-model storyteller for tension-rich RP.
z-ai/glm-5.1
Z.AI's generational leap for sustained, self-refining RP across the longest narrative arcs.
nousresearch/hermes-4-405b
Premium powerhouse with 405B parameters for intelligent RP.
anthropic/claude-haiku-4.5
Lightning-fast premium intelligence with extended thinking capabilities.
z-ai/glm-5.3
Z.AI's sharpest 1M-context reasoner for long-horizon, self-consistent RP.
x-ai/grok-4.7
Grok 4.7 brings configurable reasoning to long-context RP.
qwen/qwen3.8-max
Alibaba's Qwen 3.8 flagship for million-token, reasoning-rich multimodal RP.
x-ai/grok-4.6
xAI's Grok 4.6 flagship for sharp, coherent, long-context RP.
moonshotai/kimi-k3
Moonshot AI's 2.8T multimodal flagship for million-token, long-horizon RP.
aion-labs/aion-3.0
AionLabs' GLM-based collaborative model for high-tension storytelling.
anthropic/claude-sonnet-4.6
Anthropic's most capable Sonnet for premium coding, agentic workflows, and professional-grade RP.
Older models, still fully usable, kept for compatibility and preference.
deepseek/deepseek-v4-flash-0731
A refreshed V4 Flash with sharper reasoning for fast, coherent long-context RP.
deepseek/deepseek-chat-v3-0324
The legacy MoE model that delivers premium-quality narrative depth to your RP experiences.
deepseek/deepseek-v3.2
Experimental architecture with sparse attention for efficient long-context RP and reasoning tasks.
deepseek/deepseek-v3.1-terminus
A popular model with hybrid reasoning and advanced RP capabilities.
minimax/minimax-m2.7
Premium intelligence with advanced ESMS for exceptional storytelling.
qwen/qwen3.7-plus
Qwen's cost-efficient multimodal upgrade for million-token, reasoning-rich RP.
qwen/qwen3.6-plus
SFW premium intelligence with million-token memory for deep, wholesome RP.
moonshotai/kimi-k2.5
Moonshot AI's flagship model for roleplay with advanced reasoning capabilities.
z-ai/glm-4.6
The breakout newcomer exceeding every expectation, and maybe the new best
deepseek/deepseek-r1-0528
The open-source giant that reasons through complex plots for sophisticated RP narratives.
google/gemini-3-flash-preview
The ultimate SFW roleplay experience with premium-level intelligence and million-token context.
z-ai/glm-5.2
Z.AI's 1M-context flagship for long-horizon, self-consistent RP at massive scale.
moonshotai/kimi-k2.6
Moonshot AI's next-generation multimodal model for long-horizon coding and agent orchestration.
aion-labs/aion-2.0
A DeepSeek variant shaped for RP, built to deliver tension, conflict, and dark narrative depth.
x-ai/grok-4.3
xAI's Grok 4.3 flagship for precise, coherent, long-context RP.
anthropic/claude-sonnet-4.5
The ultimate agentic model for extended autonomous operation and advanced coding workflows.