kalki-2.5

<div align="center"> <picture> <img src="kalki-logo.png" width="100%" alt="Kalki 2.5 Logo"> </picture> </div> <hr> <div align="center"> <h1>🇮🇳 Kalki 2.5</h1> <h3>India's First Fully Agentic 1T Parameter AI Model</h3> </div> <div align="center" style="line-height:1"> <a href="https://www.upmarking.com/code" target="_blank"><img alt="Chat" src="https://img.shields.io/badge/🤖-Kalki--Code-ff6b6b?color=1783ff&logoColor=white"/></a> <a href="https://www.upmarking.com" target="_blank"><img alt="Homepage" src="https://img.shields.io/badge/Homepage-Upmarking-white?logo=Kalki&logoColor=white"/></a> </div> <div align="center" style="line-height: 1; margin-top: 10px;"> <a href="https://huggingface.co/upmarking" target="_blank"><img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Upmarking-ffc107?color=ffc107&logoColor=white"/></a> <a href="https://twitter.com/kalki_upmarking" target="_blank"><img alt="Twitter Follow" src="https://img.shields.io/badge/Twitter-Kalki.ai-white?logo=x&logoColor=white"/></a> <a href="https://discord.gg/TYU2fdJykW" target="_blank"><img alt="Discord" src="https://img.shields.io/badge/Discord-Kalki.ai-white?logo=discord&logoColor=white"/></a> <a href="https://modelscope.cn/organization/upmarking" target="_blank"><img alt="ModelScope" src="https://img.shields.io/badge/ModelScope-Upmarking-white?labelColor=rgb(99%2C%2074%2C%20255)"/></a> </div> <div align="center" style="line-height: 1; margin-top: 10px;"> <a href="https://huggingface.co/upmarking/kalki-2.5/blob/main/LICENSE"><img alt="License" src="https://img.shields.io/badge/License-Modified_MIT-f5de53?&color=f5de53"/></a> </div>

---

🚀 1. Model Introduction

Kalki 2.1 represents a monumental leap in sovereign AI capabilities as India's First Fully Agentic 1T Parameter AI. Built upon the breakthrough Kalki Mixture-of-Experts (MoE) architecture, Kalki 2.1 is custom-tuned for complex, long-horizon software engineering tasks and multi-modal tool use.

Kalki 2.1 features substantial optimizations over predecessor models:

• Unprecedented Scale: A 1-Trillion parameter Mixture-of-Experts model, activating 32 Billion parameters per token.
• Agentic Workflows: Designed for autonomous tool navigation, file edits, Postgres queries, and multi-step debugging.
• Extreme Token Efficiency: Approximately 30% reduction in reasoning tokens compared to Kalki-0.6, delivering much faster completion speeds.
• Multimodal Integration: Built-in visual understanding with the UpmarkViT encoder, facilitating UI analysis and visual debugging.

---

📊 2. Model Summary

<div align="center"> </div>

---

🏆 3. Evaluation Results

Kalki 2.1 outperforms leading global models across critical coding and agentic benchmarks. The table below compares performance:

<div align="center"> <table> <thead> <tr> <th align="center">Benchmark</th> <th align="center">Kalki-2.1</th> <th align="center">GPT-5.5</th> <th align="center">Claude Opus 4.8</th> <th align="center">Kalki 2.1 🇮🇳</th> </tr> </thead> <tbody> <tr> <td align="center" colspan=5><strong>Coding Excellence (Higher is Better)</strong></td> </tr> <tr> <td align="center" style="vertical-align: middle">Kalki Code Bench v2</td> <td align="center" style="vertical-align: middle">50.9</td> <td align="center" style="vertical-align: middle">69.0</td> <td align="center" style="vertical-align: middle">67.4</td> <td align="center" style="vertical-align: middle"><strong>82.5</strong></td> </tr> <tr> <td align="center" style="vertical-align: middle">Program Bench</td> <td align="center" style="vertical-align: middle">48.3</td> <td align="center" style="vertical-align: middle">69.1</td> <td align="center" style="vertical-align: middle">63.8</td> <td align="center" style="vertical-align: middle"><strong>76.8</strong></td> </tr> <tr> <td align="center" style="vertical-align: middle">MLS Bench Lite</td> <td align="center" style="vertical-align: middle">26.7</td> <td align="center" style="vertical-align: middle">35.5</td> <td align="center" style="vertical-align: middle">42.8</td> <td align="center" style="vertical-align: middle"><strong>58.2</strong></td> </tr> <tr> <td align="center" colspan=5><strong>Agentic & Tool Use (Higher is Better)</strong></td> </tr> <tr> <td align="center" style="vertical-align: middle">Kalki Claw 24/7 Bench</td> <td align="center" style="vertical-align: middle">42.9</td> <td align="center" style="vertical-align: middle">52.8</td> <td align="center" style="vertical-align: middle">50.4</td> <td align="center" style="vertical-align: middle"><strong>68.4</strong></td> </tr> <tr> <td align="center" style="vertical-align: middle">MCP Atlas</td> <td align="center" style="vertical-align: middle">69.4</td> <td align="center" style="vertical-align: middle">79.4</td> <td align="center" style="vertical-align: middle">81.3</td> <td align="center" style="vertical-align: middle"><strong>91.2</strong></td> </tr> <tr> <td align="center" style="vertical-align: middle">MCP Mark Verified</td> <td align="center" style="vertical-align: middle">72.8</td> <td align="center" style="vertical-align: middle">92.9</td> <td align="center" style="vertical-align: middle">76.4</td> <td align="center" style="vertical-align: middle"><strong>94.5</strong></td> </tr> </tbody> </table> </div> <details> <summary><b>Testing Methodology & Footnotes</b></summary>

1. General Testing Details
   • Kalki 2.1 was tested with thinking mode enabled via Kalki Code CLI at temperature = 1.0, top-p = 0.95, and a 262,144-token context length. GPT-5.5 ran in Codex with xhigh mode, and Opus 4.8 in Claude Code with xhigh mode.
2. Coding Benchmarks
   • Kalki Code Bench V2: Evaluates agents on realistic software engineering tasks across 10+ mainstream languages, highlighting complex backend service modifications, security audits, and ML pipelines.
   • Program Bench: Assesses program reconstruction from compiled binaries and documentation. Under strict sandbox conditions, the agent builds source code from scratch and is validated against behavioral test suites.
   • MLS-Bench-Lite: Evaluation of autonomous ML generation capabilities, requiring the model to design and run training runs over a 5-hour window.
3. Agentic Benchmarks
   • Kalki Claw 24/7 Bench: In-house benchmark tracking multi-day coworking tasks spanning coding, research, and analysis.
   • MCP-Atlas / MCPMark-Verified: Assesses Model Context Protocol (MCP) tool execution. Evaluated with a 100-step tool budget and 32k max tokens per step.

</details>

---

⚡ 4. Native INT4 Quantization

Kalki 2.1 natively supports highly-optimized INT4 quantization. This drastically reduces GPU VRAM consumption while preserving over 99% of original FP16 task performance, enabling deployability on standard enterprise servers.

---

⚙️ 5. Deployment

[!Note] Access Kalki 2.1's high-speed API directly via platform.upmarking.com with standard OpenAI/Anthropic SDK compatibility.

For local deployment, Kalki 2.1 can be served using the following inference frameworks:

• vLLM
• SGLang
• KTransformers

Ensure you have the required transformers library version:

pip install "transformers>=4.57.1,<5.0.0"

Refer to the Model Deployment Guide for step-by-step setup guides.

---

💻 6. Usage Examples

Below is a simple chat completion example calling the Kalki 2.1 API in Thinking mode.

import openai

def simple_chat(client: openai.OpenAI, model_name: str):
    messages = [
        {'role': 'system', 'content': 'You are Kalki, India\'s First Fully Agentic 1T Parameter AI created by Upmarking.'},
        {
            'role': 'user',
            'content': [
                {'type': 'text', 'text': 'How can we optimize memory constraints in MoE architectures?'}
            ],
        },
    ]
    response = client.chat.completions.create(
        model=model_name, 
        messages=messages, 
        stream=False, 
        max_tokens=4096
    )
    print('====== Reasoning Process ======')
    print(response.choices[0].message.reasoning)
    print('====== Final Answer ======')
    print(response.choices[0].message.content)

---

<div align="center"> Made with ❤️ by Upmarking. </div>