Prism ML API
Access Prism ML instantly with Puter.js, and add AI to any app in a few lines of code without backend or API keys.
// npm install @heyputer/puter.js
import { puter } from '@heyputer/puter.js';
puter.ai.chat("Explain AI like I'm five!", {
model: "prism-ml/ternary-bonsai-27b"
}).then(response => {
console.log(response);
});
<html>
<body>
<script src="https://js.puter.com/v2/"></script>
<script>
puter.ai.chat("Explain AI like I'm five!", {
model: "prism-ml/ternary-bonsai-27b"
}).then(response => {
console.log(response);
});
</script>
</body>
</html>
List of Prism ML Models
Ternary Bonsai 2 27B
prism-ml/ternary-bonsai-2-27b
Ternary Bonsai 2 27B is PrismML's second-generation ternary-weight compression of a Qwen3.8 27B base model, following the original Bonsai 27B built on Qwen3.6. Weights are stored in a ternary {-1, 0, +1} format with FP16 group-wise scaling, giving roughly 1.76 effective bits per weight and a 5.9GB footprint, about 9x smaller than the full-precision model. Across a 20-benchmark thinking-mode suite, it averages 83.9, retaining 98.2% of the Qwen3.8 27B baseline's 85.4 score, up from the roughly 95% retention PrismML reported for the first-generation Bonsai 27B. Category scores include 96.57 on math and 81.58 on coding. It supports a 262K-token context window, tool calling, and image input alongside text, making it suited for coding agents, math-heavy workloads, and long-document or multimodal tasks where compute budget matters.
ChatTernary Bonsai 27B
prism-ml/ternary-bonsai-27b
Ternary Bonsai 27B is PrismML's ternary-weight build of Qwen3.6 27B, compressing the model to roughly 1.71 bits per weight while retaining about 95% of full-precision quality. Across a 15-benchmark suite covering knowledge, reasoning, math, coding, instruction following, tool calling, and vision, it averages 80.49 in thinking mode, including 99.20 on MATH-500, 93.90 on HumanEval+, 82.75 on LiveCodeBench, and 74.41 on BFCL v3 for tool calling. It also handles vision input, scoring 68.96 on MMMU-Pro. The model supports a 262K-token context window and structured JSON output, making it suited for long-document analysis, coding agents, and tool-calling workflows. PrismML also offers a 1-bit variant of Bonsai 27B with a smaller footprint; the ternary build trades some of that compression for higher accuracy across the same benchmark suite.
Frequently Asked Questions
The Prism ML API gives you access to models for AI chat. Through Puter.js, you can start using Prism ML models instantly with zero setup or configuration.
Puter.js supports a variety of Prism ML models, including Ternary Bonsai 2 27B and Ternary Bonsai 27B. Find all AI models supported by Puter.js in the AI model list.
With the User-Pays model, users cover their own AI costs through their Puter account. This means you can build apps without worrying about infrastructure expenses.
Puter.js is a JavaScript library that provides access to AI, storage, and other cloud services directly from a single API. It handles authentication, infrastructure, and scaling so you can focus on building your app.
Yes — the Prism ML API through Puter.js works with any JavaScript framework, Node.js, or plain HTML. Just include the library and start building. See the documentation for more details.