Kingy AI Launch Intelligence
Kingy AI tracks AI models by provider, model family, modality, release status, API access, web access, local availability, open-weight/open-source status, pricing notes, hardware requirements, official sources, benchmark caveats, and last-verified status.
These paths use the same source-backed filters as the directory. They are discovery routes, not rankings.
Showing 25–48 of 51 models
Source-checked profiles within the review window appear first; retired models appear last. Check the date and limitations on each record. Indexing is assessed separately from review age.
Llama-3.1-Nemotron-Nano-8B-v1 is an NVIDIA compact open model that fits on a single RTX GPU and supports 128K context.
Llama-3.3-Nemotron-Super-49B-v1.5 is an NVIDIA reasoning model derived from Meta Llama 3.3 and tuned for RAG and tool calling.
Ministral 3 14B is a Mistral small-model variant listed with the Ministral 3 family in official documentation.
NVIDIA Nemotron 3 Super 120B-A12B is part of NVIDIA's Nemotron open-model family for agentic AI and reasoning workflows.
Qwen3-30B-A3B is a Qwen3 mixture-of-experts open-weight model described in Qwen's official Qwen3 release.
Claude Haiku 4.5 is Anthropic's Haiku-tier model described in official docs as the fastest current model with near-frontier intelligence.
DeepSeek V4 Flash is listed in DeepSeek API docs as a current model supporting thinking and non-thinking modes, JSON output, tool calls, and a 1M context length.
Devstral 2 is listed by Mistral as a frontier code agents model for software engineering tasks.
Gemini 3 Flash is a Google Gemini API preview model positioned as frontier-class performance at a lower cost tier than larger models.
Gemini 3.1 Pro is a Google Gemini API preview model described for advanced intelligence, complex problem-solving, and agentic or vibe-coding capabilities.
Gemini 3.5 Flash is a Google Gemini API model listed as stable and positioned for sustained frontier performance on agentic and coding tasks.
GPT-5.4 is an OpenAI API model described in official docs as a more affordable model for coding and professional work.
GPT-5.4 mini is an OpenAI API model described as a smaller model for coding, computer use, and subagents.
GPT-5.5 is OpenAI's flagship API model for complex reasoning, coding, and professional work, with text and image input and text output described in official OpenAI model docs.
Grok 4.3 is an xAI API model listed with agentic tool calling, configurable reasoning, and a 1M-token context window.
Grok Build 0.1 is an xAI coding model described as trained specifically for fast agentic coding workflows.
Llama 4 Maverick is a larger Meta Llama 4 mixture-of-experts model optimized for multimodal understanding, coding, tool-calling, and agentic systems.
Llama 4 Scout is a Meta Llama 4 mixture-of-experts model optimized for multimodal understanding, multilingual tasks, coding, tool-calling, and agentic systems.
Qwen3-235B-A22B is a flagship Qwen3 mixture-of-experts model described in Qwen release materials as open-weight and useful across coding, math, general capabilities, and agent workflows.
Claude Opus 4.6 is an Anthropic Opus-tier model included in Claude pricing, data residency, and long-context notes.
Claude Opus 4.7 is an Anthropic Opus-tier model listed in official Claude pricing and lifecycle documentation.
Claude Sonnet 4.5 is a Sonnet-tier Claude model listed by Anthropic with current pricing and cloud-platform notes.
DeepSeek-V3 is a DeepSeek model release described by DeepSeek as improving speed and API compatibility.
DeepSeek-V3.2 is a DeepSeek model release described by DeepSeek as integrating thinking into tool-use.