InclusionAI Ling and Ring 2.6 Models Are Now Available in Puter.js

Puter.js now supports the full Ling and Ring 2.6 lineup from InclusionAI (Ant Group) — a high-efficiency Flash tier, a trillion-parameter fast-thinking flagship, and a dedicated reasoning model purpose-built for autonomous agents.

What is Ling 2.6 Flash?

Ling 2.6 Flash is the efficiency-focused tier in the Ling family — a Mixture-of-Experts model with 104B total parameters and only 7.4B active per token, delivering strong performance at a fraction of the compute cost.

• 262K context window with up to 32K output tokens
• Intelligence Index of 26 on Artificial Analysis — nearly double the open-weight non-reasoning median of 13, and a 10-point jump over Ling-flash-2.0
• 59.3% on GPQA Diamond
• Trained with Agentic Reinforcement Learning for tool use, terminal operations, and multi-step agent tasks
• Well suited for cost-sensitive agent pipelines and high-throughput automation

What is Ling 2.6 1T?

Ling 2.6 1T is InclusionAI's trillion-parameter non-reasoning flagship, built by Ant Group's AGI initiative. It uses a Mixture-of-Experts architecture with roughly 50B active parameters per token, employing a "fast thinking" approach that cuts token costs to about a quarter of comparable models while keeping flagship-tier output quality.

• 262K context window with up to 32K output tokens
• Intelligence Index of 34 on Artificial Analysis — far above the open-weight non-reasoning median of 13
• State-of-the-art on AIME 2025 and SWE-bench Verified, and #1 among open-source models on ArtifactsBench for front-end code generation
• Flagship-level reasoning without a dedicated thinking model
• Strong fit for autonomous agents, large-scale coding workflows, and cost-sensitive pipelines

What is Ring 2.6 1T?

Ring 2.6 1T is the reasoning counterpart — a trillion-parameter MoE with 63B active parameters per token, purpose-built for agentic workflows like coding agents, tool use, and long-horizon multi-step execution. It introduces adaptive reasoning effort with high and xhigh modes that dynamically allocate thinking budget based on task complexity.

• 262K context window with up to 66K output tokens
• 95.83 on AIME 2026 and 88.27 on GPQA Diamond
• 87.60 on PinchBench in agent mode — ahead of GPT-5.4 and Gemini 3.1 Pro
• Leading scores on ClawEval, TAU2-Bench, and GAIA2-search
• A strong pick for autonomous agent systems and reasoning-heavy pipelines

Examples

High-throughput agent work with Ling 2.6 Flash

puter.ai.chat("Write a Node.js script that watches a directory, parses any new JSON files, validates them against a schema, and moves invalid files to a quarantine folder", {
    model: "inclusionai/ling-2.6-flash"
})

Fast-thinking flagship with Ling 2.6 1T

puter.ai.chat("Refactor this Express handler into a typed Fastify route with input validation, structured logging, and proper error responses", {
    model: "inclusionai/ling-2.6-1t"
})

Long-horizon reasoning with Ring 2.6 1T

puter.ai.chat("Design an autonomous agent that monitors a Postgres database for anomalies, opens a triage ticket, and proposes a remediation plan. Walk through the architecture and failure modes.", {
    model: "inclusionai/ring-2.6-1t"
})

Agentic coding across a long context

puter.ai.chat(`Given this repository tree and these files, identify where authentication
state is leaking between requests and propose a minimal fix:\n\n${repoDump}`, {
    model: "inclusionai/ring-2.6-1t"
})

Streaming with reasoning traces

const response = await puter.ai.chat(
    "Prove that the sum of the first n odd numbers equals n², then generalize the result",
    { model: "inclusionai/ring-2.6-1t", stream: true }
);

for await (const part of response) {
    if (part?.reasoning) puter.print(part?.reasoning);
    else puter.print(part?.text);
}

Get Started Now

Just add one library to your project:

// npm install @heyputer/puter.js
import { puter } from '@heyputer/puter.js';

Or add one script tag to your HTML:

<script src="https://js.puter.com/v2/"></script>

No API keys or account needed. Start building with Ling 2.6 Flash, Ling 2.6 1T, and Ring 2.6 1T immediately.

Learn more:

• Puter.js AI Documentation