Model Library

GLM 5.3 Flash

GLM

Available//01

Long-context chat, coding and tool workflows with GLM 5.3 Flash.

• Chat
• Vision
• Video

Parameters

320B / 18B active

Context

1,048,576 tokens

Max output

131,072 tokens

Rate / 1M

$0.125 in · $0.050 cached · $0.50 out

DeepSeek V4.1 Flash

DeepSeek

Coming soon//01

Multimodal mixture-of-experts model: 552B backbone parameters, with 8B active during prefill and 16B during decoding.

• Chat
• Vision

Parameters

552B / 8B prefill / 16B decode active

Context

1,048,576 tokens

Qwen 3.6 35B A3B

Qwen

Coming soon//02

Vision-language mixture-of-experts model: 35B total, 3B active parameters.

• Chat

Parameters

35B / 3B active

Context

262,144 tokens

Output guidance

81,920 tokens

Kimi K3

Kimi

Roadmap//01

Open-weight mixture-of-experts model: 2.8T total, 104B active parameters.

• Chat

Parameters

2.8T / 104B active

Context

1,048,576 tokens

GLM 5.3

GLM

Roadmap//02

A longer-term GLM family target for the Arnict catalog.

• Chat

Qwen 3.8 2.4T A95B

Qwen

Roadmap//03

Text-only mixture-of-experts model: 2.4T total, 95B active parameters.

• Chat
• Reasoning

Parameters

2.4T / 95B active

Context

262,144 tokens

Output guidance

262,144 tokens

DeepSeek V4 Pro 0813

DeepSeek

Roadmap//04

DeepSeek V4 Pro mixture-of-experts model with DSpark speculative decoding and a 1,048,576-token context.

• Chat
• Reasoning

Context

1,048,576 tokens

Output guidance

384,000 tokens