Kimi Linear 48B A3B Instruct
Intelligence
14.4
#353/520
Coding
14.2
#270/426
Math
36.3
#171/265
Speed
—
Pricing
—
Kimi Linear 48B A3B Instruct is Kimi’s advanced AI model designed for various applications in natural language processing. With a Math Index of 36.3, it excels in mathematical tasks, making it suitable for users requiring strong analytical capabilities.
When to Use Kimi Linear 48B A3B Instruct
✓ Best For
- Complex mathematical problem solving.
- Data analysis and interpretation.
- Educational tools for teaching mathematics.
✗ Not Ideal For
- General conversational AI tasks.
- Applications requiring high-speed processing.
How Kimi Linear 48B A3B Instruct Compares
Intelligence Index · Higher is better
MistralAlibabaKimiMBZUAI Institute of Foundation ModelsNVIDIA
Benchmark Profile
Coding Index
AlibabaKimiOpenAIMiniMax
Math Index
MistralAnthropicKimiMBZUAI Institute of Foundation ModelsGoogle
Intelligence · Coding · Math
Intelligence
Coding
Math
All Benchmark Scores (12)
| Benchmark | Score | |
|---|---|---|
| Intelligence Index | 14.4 | |
| Coding Index | 14.2 | |
| Math Index | 36.3 | |
| MMLU-Pro | 585% | |
| GPQA | 412% | |
| LiveCodeBench | 378% | |
| HLE | 27% | |
| SciCode | 19.9% | |
| IFBench | 28.1% | |
| LCR | 25.7% | |
| TerminalBench Hard | 11.4% | |
| AIME 2025 | 36.3% |
Data: Artificial Analysis · Updated: March 26, 2026
Frequently Asked Questions (15)
When was Kimi Linear 48B A3B Instruct released?
Kimi Linear 48B A3B Instruct was released on October 30, 2025.
Who created Kimi Linear 48B A3B Instruct?
Kimi Linear 48B A3B Instruct was created by Kimi.
How intelligent is Kimi Linear 48B A3B Instruct?
Kimi Linear 48B A3B Instruct scores 14 (estimated) on the Artificial Analysis Intelligence Index, placing it above average among other open weight non-reasoning models of similar size (median: 13).
How verbose is Kimi Linear 48B A3B Instruct?
When evaluated on the Intelligence Index, Kimi Linear 48B A3B Instruct generated 32M output tokens, which is at the higher end compared to other open weight non-reasoning models of similar size (median: 3.8M).
Is Kimi Linear 48B A3B Instruct a reasoning model?
No, Kimi Linear 48B A3B Instruct is not a reasoning model. It provides direct responses without extended chain-of-thought reasoning.
What input modalities does Kimi Linear 48B A3B Instruct support?
Kimi Linear 48B A3B Instruct supports text input.
What output modalities does Kimi Linear 48B A3B Instruct support?
Kimi Linear 48B A3B Instruct supports text output.
Can Kimi Linear 48B A3B Instruct process images?
No, Kimi Linear 48B A3B Instruct does not support image input. It can only process text.
Is Kimi Linear 48B A3B Instruct multimodal?
No, Kimi Linear 48B A3B Instruct is not multimodal. It only supports text input.
What is the context window of Kimi Linear 48B A3B Instruct?
Kimi Linear 48B A3B Instruct has a context window of 1.0M tokens. This determines how much text and conversation history the model can process in a single request.
Is Kimi Linear 48B A3B Instruct open source?
Yes, Kimi Linear 48B A3B Instruct is open weights. The model weights are publicly available and can be downloaded for self-hosting.
How many parameters does Kimi Linear 48B A3B Instruct have?
Kimi Linear 48B A3B Instruct has 49.1 billion parameters (3 billion active).
What are the active parameters of Kimi Linear 48B A3B Instruct?
Kimi Linear 48B A3B Instruct is a Mixture of Experts (MoE) model with 49.1 billion total parameters, but only 3 billion active parameters are used during inference.
What is the license for Kimi Linear 48B A3B Instruct?
Kimi Linear 48B A3B Instruct is released under the MIT license. This license allows commercial use.
How does Kimi Linear 48B A3B Instruct perform on benchmarks?
Kimi Linear 48B A3B Instruct achieves a score of 14 on the Artificial Analysis Intelligence Index. This composite benchmark evaluates models across reasoning, knowledge, mathematics, and coding.
