Available Models

FreeInference provides access to multiple state-of-the-art LLM models for coding agents and IDEs.

Model Overview

Model ID

Name

Context Length

Max Output

Features

glm-5.1

GLM-5.1

200K tokens

128K tokens

Function calling, Structured output, Bilingual (Chinese/English), Thinking mode, Tool streaming

glm-5.2

GLM-5.2

1M tokens

128K tokens

Function calling, Structured output, Bilingual (Chinese/English), Thinking mode, Tool streaming, Internal/staff access only

glm-5-turbo

GLM-5 Turbo

200K tokens

128K tokens

Function calling, Structured output, Bilingual (Chinese/English), Thinking mode, Tool streaming

qwen3.6-35b

Qwen3.6 35B

262K tokens

8K tokens

Function calling, Structured output

minimax-m2.5

MiniMax M2.5

205K tokens

131K tokens

Function calling, Structured output, Thinking mode

minimax-m2.7

MiniMax M2.7

205K tokens

131K tokens

Function calling, Structured output, Multimodal (text+image)

minimax-m3

MiniMax M3

1M tokens

131K tokens

Function calling, Structured output, Thinking mode, Multimodal (text+image)

kimi-k2.7-code

Kimi K2.7 Code

262K tokens

131K tokens

Function calling, Structured output, Thinking mode, Coding agents only

Model Details

GLM-5.1

Model ID: glm-5.1

Aliases: freeinference-glm-5.1

  • Context length: 200,000 tokens

  • Max output: 128,000 tokens

  • Quantization: fp8

  • Input modalities: text

  • Output modalities: text

  • Language support: Chinese, English

  • Function calling: Yes

  • Structured output: Yes

  • Thinking mode: Yes

  • Tool streaming: Yes

GLM-5.2

Model ID: glm-5.2

Aliases: freeinference-glm-5.2

Internal/staff access only (required_role: internal) — a public API key gets 403.

  • Context length: 1,000,000 tokens

  • Max output: 128,000 tokens

  • Quantization: fp8

  • Input modalities: text

  • Output modalities: text

  • Language support: Chinese, English

  • Function calling: Yes

  • Structured output: Yes

  • Thinking mode: Yes

  • Tool streaming: Yes

GLM-5 Turbo

Model ID: glm-5-turbo

Aliases: freeinference-glm-5-turbo

  • Context length: 200,000 tokens

  • Max output: 128,000 tokens

  • Quantization: fp8

  • Input modalities: text

  • Output modalities: text

  • Language support: Chinese, English

  • Function calling: Yes

  • Structured output: Yes

  • Thinking mode: Yes

  • Tool streaming: Yes

Qwen3.6 35B

Model ID: qwen3.6-35b

  • Context length: 262,144 tokens

  • Max output: 8,192 tokens

  • Quantization: fp8

  • Input modalities: text

  • Output modalities: text

  • Function calling: Yes

  • Structured output: Yes

MiniMax M2.5

Model ID: minimax-m2.5

  • Context length: 204,800 tokens

  • Max output: 131,072 tokens

  • Quantization: bf16

  • Input modalities: text

  • Output modalities: text

  • Function calling: Yes

  • Structured output: Yes

  • Thinking mode: Yes

MiniMax M2.7

Model ID: minimax-m2.7

  • Context length: 204,800 tokens

  • Max output: 131,072 tokens

  • Quantization: bf16

  • Input modalities: text, image

  • Output modalities: text

  • Function calling: Yes

  • Structured output: Yes

MiniMax M3

Model ID: minimax-m3

  • Context length: 1,048,576 tokens

  • Max output: 131,072 tokens

  • Quantization: bf16

  • Input modalities: text, image

  • Output modalities: text

  • Function calling: Yes

  • Structured output: Yes

  • Thinking mode: Yes

Kimi K2.7 Code

Model ID: kimi-k2.7-code

Aliases: Kimi-K2.7-Code, kimi-k2.7

Served through the Kimi Code coding-plan subscription; intended for coding agents rather than general chat.

  • Context length: 262,144 tokens

  • Max output: 131,072 tokens

  • Quantization: fp8

  • Input modalities: text

  • Output modalities: text

  • Function calling: Yes

  • Structured output: Yes

  • Thinking mode: Yes

Switching Models

To use different models, change the model name in your IDE configuration:

Cursor: Select from the dropdown in settings

Kilo Code: Select from the dropdown in extension settings. A good default is glm-5.1; switch to glm-5-turbo for faster iteration, minimax-m3 for long-context and image-aware workflows, or kimi-k2.7-code for agentic coding.

Roo Code: Select from the dropdown in extension settings