Nex N2.6

LatestLanguage model

Nex N2.6 is post-trained from GLM-5.3-Flash and served by MOSS. Use it for text conversations, image and video understanding, and client-side function calling. The generation endpoints support synchronous and streaming text output. Every valid MOSS API key can call Nex N2.6.

View API reference

ENDPOINT

POST /v1/chat/completions

Current request path

Input

Text, images, and video

Content accepted by the request

Output

Text

Content returned on success

Model overview

Use the nex-n2.6 alias for routine calls. Use a dated snapshot to pin a version.

Billing

Free for a limited time.

The response still reports token usage in usage.

0 credits View full pricing

Modality support

TTextInput and output

AAudioNot supported

IImageInput only

VVideoInput only

Images: Chat Completions, Responses, and Messages all support images. Use the image content block for your protocol. Video: Chat Completions and Responses accept MP4 and WebM through video_url. Messages does not support video. Files, file_id, and asset library references are currently unsupported. Public media URLs must be accessible to MOSS. The entire request body, including encoded media, must be at most 10 MiB.

Supported endpoints

Chat Completions Standard text conversations or clients using OpenAI Chat Completions./v1/chat/completionsSupported

Responses Use the OpenAI Responses format and follow the Nex N2.6 input rules./v1/responsesSupported

Messages Use the Anthropic Messages format./v1/messagesSupported

Count input tokens Estimate usage for text-only input before a Messages generation request./v1/messages/count_tokensSupported

Single-speaker speech /v1/audio/speechNot supported

Multi-speaker dialogue speech /v1/audio/speech/speakersNot supported

Audio transcription /v1/audio/transcriptionsNot supported

Voice design /v1/audio/voice/generationsNot supported

Realtime interaction /v1/realtimeNot supported

Image generation /v1/images/generationsNot supported

Chat Completions and Responses use Bearer authentication. Messages and Count Tokens use x-api-key and require anthropic-version: 2023-06-01. The production HTTP origin is https://api.mosi.cn. Set the OpenAI SDK base_url to https://api.mosi.cn/v1.

Capabilities

Performance—

Inference speed—

Streaming Supported

Async task API Not supported

Streaming output: All three generation APIs support streaming. Read deltas and termination events according to the selected protocol.
Multi-turn conversations: The client sends conversation history with each request. Responses is stateless.
Function calling: With Chat Completions and Messages, the client can execute functions and send the results back.
Output budget: The production default is 8,192 tokens and the maximum setting is 131,072 tokens. Use max_tokens for Chat Completions and Messages, and max_output_tokens for Responses.

Versions and snapshots

Model IDnex-n2.6

Snapshotnex-n2.6-2026-11-05Latest

The model field in the generation response identifies the snapshot used, nex-n2.6-2026-11-05. You can omit version. If you specify it, use 2026-11-05 and make sure it matches the snapshot selected by model.