Nexith Core は自社主権 GPU クラスタで稼働開始 —— キャッシュヒットは最大90%オフ、OpenAI 互換。詳細を見る
Nexith
登録
モデル

ひとつの基盤。
広がるモデルファミリー。

Nexith Core は推論、文章作成、コードのための汎用モデルとして提供中です。統合ルーティングと画像生成も同じ基盤とOpenAI互換APIで順次提供します。

Available now

Nexith Core

NX-8B
Language & reasoning

Flagship reasoning, writing, and code. 1M context, text + image in, over an OpenAI-compatible API.

Nexith Coreを見る
Nexith Core
Refactor this module and explain the tradeoffs.
I split the parser into three passes and kept the hot path allocation-free. Two tradeoffs:
1.Streaming keeps memory flat but adds a small latency floor.
2.The cache trades freshness for up to 90% off on repeats.
1M context~119 tok/s
Preview

Nexith

NX-ONE
Unified routing

One endpoint that routes each request to the right capability — automatically.

詳しく見る
Nexith · one endpoint
POST /v1/chat
Reads the request
Works out what it needs, then routes.
ReasoningNexith Core
ImageNexith Image
Audiosoon
Videosoon
Coming soon

Nexith Image

NX-I1
Image generation

Generate and edit images from a prompt — coming to Apps and the API.

詳しく見る
Nexith Image
A calm studio still-life, soft morning light…
Text → ImageEditing

One endpoint. A family that keeps growing.

Start with Core today — the API is the one you already know. As unified routing and image generation land, they show up at the same endpoint. Up to 90% off on cache hits, streaming everywhere, no seats or minimums.

Get started freeSee pricing
Nexith — フロンティア AI モデル | Nexith