Help CentreFeaturesAvailable Chat Models
Features

Chat Models (96 total)

Premium tier (66): Best quality, advanced reasoning

Fast tier (30): Quick responses, cost-effective

OpenAI:

  • •GPT-5.6 — Flagship GPT-5.6 alias — routes to GPT-5.6 Sol for complex reasoning and coding [premium, vision, reasoning]
  • •GPT-5.6 Sol — Frontier GPT-5.6 model for complex professional work, coding, and agentic tasks [premium, vision, reasoning]
  • •GPT-5.6 Terra — GPT-5.6 model balancing intelligence and cost — strong agentic execution [premium, vision, reasoning]
  • •GPT-5.6 Luna — Cost-efficient GPT-5.6 model for high-volume workloads and subagent loops [fast, vision, reasoning]
  • •GPT-5.5 — Prior frontier model — prefer GPT-5.6 Sol for new projects [premium, vision, reasoning]
  • •GPT-5.4 — Highly capable GPT model for coding and agentic tasks — prefer GPT-5.6 Terra at same price [premium, vision, reasoning]
  • •GPT-5.4 mini — Fast GPT-5.4 variant for coding, computer use, and high-volume subagent workloads [fast, vision, reasoning]
  • •GPT-5.4 Pro — Most powerful GPT model with maximum compute for complex reasoning [premium, vision, reasoning]
  • •GPT-5.2 — The best model for coding and agentic tasks across industries [premium, vision, reasoning]
  • •GPT-5.1 — Intelligent reasoning model with configurable reasoning effort [premium, vision, reasoning]
  • •GPT-5 — Previous intelligent reasoning model for coding and agentic tasks [premium, vision, reasoning]
  • •GPT-5 mini — A faster, cost-efficient version of GPT-5 [fast, vision]
  • •GPT-5 nano — Fastest, most cost-efficient version of GPT-5 [fast, vision]
  • •GPT-5.2 Pro — Smarter and more precise responses [premium, vision, reasoning]
  • •GPT-4.1 — Smartest non-reasoning model [premium, vision]
  • •GPT-4.1 mini — Smaller, faster version of GPT-4.1 [fast, vision]
  • •o3 — Reasoning model for complex tasks [premium, reasoning]
  • •o3 Pro — Version of o3 with more compute for better responses [premium, reasoning]
  • •o4 mini — Fast, cost-efficient reasoning model [fast, reasoning]

    • Anthropic:

  • •Claude Fable 5 — Most capable widely released Claude — currently unavailable due to US export controls. Use Claude Opus 4.8 instead. [premium, vision, reasoning]
  • •Claude Opus 4.8 — Most capable Opus — enhanced coding, agentic workflows, and long-horizon reasoning with 1M context [premium, vision, reasoning]
  • •Claude Opus 4.7 — Previous Opus — enhanced SWE, vision, and long-horizon agentic reasoning with 1M context [premium, vision, reasoning]
  • •Claude Opus 4.6 — Hybrid reasoning model with 1M context, top-tier coding and agentic performance [premium, vision, reasoning]
  • •Claude Opus 4.5 — Previous premium model with maximum intelligence [premium, vision, reasoning]
  • •Claude Sonnet 5 — Best combination of speed and intelligence — adaptive thinking with near-Opus quality [premium, vision, reasoning]
  • •Claude Sonnet 4.6 — Previous Sonnet — strong speed and intelligence balance [premium, vision, reasoning]
  • •Claude Sonnet 4.5 — Previous smart model for complex agents and coding [premium, vision, reasoning]
  • •Claude Haiku 4.5 — Fastest model with near-frontier intelligence [fast, vision]

    • Google:

  • •Gemini 3.1 Pro — Most advanced reasoning model with complex problem-solving [premium, vision, reasoning]
  • •Gemini 3.1 Flash-Lite — Cheapest frontier-class model — half the cost of Gemini 3 Flash with strong tool calling [fast, vision]
  • •Gemini 3.1 Flash Live — Low-latency Live API model for real-time dialogue and voice-first AI applications [fast, vision]
  • •Gemini 3.6 Flash — Latest stable Flash — strong agentic and multimodal performance with lower output cost than 3.5 Flash [fast, vision, reasoning]
  • •Gemini 3.5 Flash-Lite — Fastest 3.5-class model — high-throughput chat, tools, and vision at lower cost than full Flash [fast, vision, reasoning]
  • •Gemini 3.5 Flash — Frontier intelligence optimized for agentic workflows, coding, and video at higher speed [fast, vision, reasoning]
  • •Gemini 3 Flash — Frontier intelligence with superior search and grounding [fast, vision, reasoning]
  • •Gemini 2.5 Pro — State-of-the-art thinking model for complex problems [premium, vision, reasoning]
  • •Gemini 2.5 Flash — Best price-performance for large scale processing [fast, vision, reasoning]
  • •Gemini 2.5 Flash-Lite — Fastest flash model for cost-efficiency [fast, vision]

    • xAI:

  • •Grok 4.5 — Frontier Grok for coding, agentic tasks, and knowledge work — tool calling, web/X search, code execution [premium, vision, reasoning]
  • •Grok 4.3 — Fast, intelligent Grok — 1M context, 3 reasoning levels, top agentic tool calling [premium, vision, reasoning]
  • •Grok 4.20 Multi-Agent — Latest Grok beta optimized for multi-agent orchestration [premium, vision, reasoning]
  • •Grok 4.20 Reasoning — Latest Grok beta with extended reasoning [premium, vision, reasoning]
  • •Grok 4.20 — Fast Grok without reasoning overhead [premium, vision]
  • •Grok 3 Mini — Smaller, faster Grok with reasoning [fast, reasoning]

    • DeepSeek:

  • •DeepSeek V4 Flash — 1M context, thinking + non-thinking modes, tool calls [fast, reasoning]
  • •DeepSeek V4 Pro — Flagship model — 1M context, thinking + non-thinking modes [premium, reasoning]

    • Alibaba (Qwen):

  • •Qwen3 Max (Direct) — Alibaba's flagship general-purpose LLM via DashScope - top-tier reasoning and coding [premium, reasoning]
  • •Qwen Deep Research — Automated deep research - plans research steps, performs web searches, generates structured reports [premium]

    • MiniMax:

  • •MiniMax M3 — Latest M-series — 1M context, agentic coding, tool use, vision + video input, interleaved thinking [premium, vision, reasoning]
  • •MiniMax M2.7 — Recursive self-improvement — SOTA in software engineering, tool calling, and office productivity [premium, reasoning]
  • •MiniMax M2.7 Highspeed — M2.7 at ~100 tps — same performance, faster and more agile [fast, reasoning]
  • •MiniMax M2.5 — Peak performance and ultimate value — master the complex [premium, reasoning]
  • •MiniMax M2.5 Highspeed — M2.5 at ~100 tps — same performance, faster and more agile [fast, reasoning]
  • •MiniMax M2.1 — Polyglot programming mastery with precision code refactoring [premium]
  • •MiniMax M2 — Agentic capabilities with function calling and advanced reasoning [premium]

    • ByteDance:

  • •Seed 2.0 Pro — ByteDance flagship — 76.5% SWE-Bench, 98.3% AIME 2025, hour-long video understanding [premium, vision, reasoning]
  • •Seed 2.0 Lite — Versatile multimodal model with low latency for agent and vision tasks [fast, vision]

    • Meta:

  • •Llama 4 Maverick — Meta's flagship Llama 4 model [premium, vision]
  • •Llama 4 Scout — Efficient Llama 4 model [fast]
  • •Llama 3.3 70B — Meta's powerful open source model [premium]
  • •Llama 3.3 70B — Meta's powerful open source model [fast]
  • •Muse Spark 1.1 — Meta multimodal agent model — text, image, video, audio, and PDF input with 1M context, parallel tool calling, and built-in search. US availability on OpenRouter. Admin preview only. [premium, vision, reasoning]

    • Mistral:

  • •Mistral Medium 3.1 — Balanced Mistral model [premium]
  • •Mistral Large 2512 — Latest large Mistral model [premium]
  • •Mistral Small Creative — Creative writing focused model [fast]

    • Qwen:

  • •Qwen3 235B — Large Qwen model with 235B parameters [premium]
  • •Qwen3 VL 235B — Vision-language Qwen model [premium, vision]
  • •Qwen3 Max — Most powerful Qwen model [premium]
  • •Qwen 3.7 Max — Flagship Qwen3.7 — 1M context, agentic coding, office productivity, long-horizon execution, prompt caching [premium, reasoning]
  • •Qwen 3.7 Plus — Cost-effective Qwen3.7 VL model — 1M context, vision + tools + agent workflows, GUI and mobile navigation [premium, vision, reasoning]

    • Z-AI:

  • •GLM 5.1 — Latest Z-AI flagship — enhanced long-horizon coding and autonomous agent tasks [premium, reasoning]
  • •GLM 5 — Z-AI flagship model with strong reasoning and tool use [premium]
  • •GLM 5 Turbo — Fast inference model optimized for agentic workflows and tool use [fast]
  • •GLM 4.7 — Latest GLM model [premium]
  • •GLM 4.6 — Powerful GLM model [premium]
  • •GLM 4.5 Air — Lightweight GLM model [fast]
  • •GLM 5.2 — Z.ai GLM 5.2 via EvoLink — long-horizon coding agents, repo Q&A, tool-using engineering workflows, 1M context with thinking + prompt caching [premium, reasoning]

    • Xiaomi:

  • •MiMo v2.5 — Native omnimodal model — Pro-level agentic performance at half the cost, with image & video understanding and 1M context [premium, vision, reasoning]
  • •MiMo v2.5 Pro — Xiaomi's 1T-parameter flagship — agentic workflows, tool calling, and advanced reasoning with 1M context [premium, reasoning]
  • •MiMo v2 Flash — Xiaomi's fast AI model [fast]

    • NVIDIA:

  • •Nemotron 3 Ultra — 550B MoE frontier reasoning model (55B active) — long-horizon agentic workflows, coding agents, and deep research with 1M context. Free on OpenRouter. Web search via Kunya smart search (native tool calling not supported). [premium, reasoning, free]
  • •Nemotron 3 Nano — Nvidia's compact free MoE model for fast chat and lightweight tasks [fast]

    • Moonshot:

  • •Kimi K3 — 2.8T MoE flagship for complex coding, knowledge work, and long-horizon agentic workflows with 1M context and native vision [premium, vision, reasoning]
  • •Kimi K2.7 Code — Coding-focused MoE model for long-horizon programming, agentic task decomposition, and multimodal reasoning [premium, vision, reasoning]
  • •Kimi K2.5 — State-of-the-art visual coding and agentic tool-calling with multimodal reasoning [premium, vision, reasoning]

    • StepFun:

  • •Step 3.5 Flash — 196B MoE reasoning model — activates 11B per token, extremely fast [fast, reasoning]
  • •Step 3.7 Flash — 196B MoE multimodal model — native image & video understanding, selectable reasoning depth [fast, vision, reasoning]

    • OpenRouter:

  • •Hunter Alpha — 1T parameter frontier model built for agentic multi-step reasoning [premium]
  • •Healer Alpha — Omni-modal frontier model with vision, hearing, reasoning, and action [premium, vision]

    • Nous Research:

  • •Hermes 4 405B — Flagship uncensored reasoning model from Nous Research — hybrid think/respond mode, low refusal rates, strong at math, code, and structured output [premium]
  • •Hermes 4 70B — Efficient uncensored reasoning model from Nous Research — hybrid think/respond mode, low refusal rates, strong at math, code, and structured output [fast]

    • Perplexity:

  • •Sonar Pro Search — Perplexity agentic search — multi-step research with citations (compare vs Brave / Gemini grounding) [premium, reasoning]

    • Tencent:

  • •Hy3 — 295B MoE reasoning model (21B active) — agentic workflows, 256K context, configurable chain-of-thought for coding, analysis, and tool use [premium, reasoning]
  • •Hy3 — 295B MoE reasoning model (21B active) — agentic workflows, 256K context, configurable chain-of-thought. Free on Kunya until July 21. [premium, reasoning, free]

    • Poolside:

  • •Laguna M.1 — Poolside flagship coding agent — tool calling, reasoning, 256K context, up to 32K output. Free on Kunya until July 28. [premium, reasoning, free]

    • Kunya:

  • •Kunya V1 — Intelligently routed model — Opus-level quality at budget cost. Routes to the best model for each request. [premium, vision, reasoning]

    • How to switch models: Use the model selector dropdown at the top of the Chat interface. You can change models at any time, even mid-conversation.

    Need more help with this topic?

    Ask Kunya AI