Qwen: Qwen3.8 2.4T A95B
qwen/qwen3.8-2.4t-a95b
Try Qwen3.8 2.4T A95B for free in your browser, and add it to your app for free with Puter.js AI API.
Try it free Add to your app// npm install @heyputer/puter.js
import { puter } from '@heyputer/puter.js';
puter.ai.chat("Explain quantum computing in simple terms", {
model: "qwen/qwen3.8-2.4t-a95b"
}).then(response => {
document.body.innerHTML = response.message.content;
});
<html>
<body>
<script src="https://js.puter.com/v2/"></script>
<script>
puter.ai.chat("Explain quantum computing in simple terms", {
model: "qwen/qwen3.8-2.4t-a95b"
}).then(response => {
document.body.innerHTML = response.message.content;
});
</script>
</body>
</html>
Model Card
Qwen3.8 2.4T A95B is Alibaba's open-weight release of its Qwen3.8 Max flagship, a sparse mixture-of-experts model with 2.4 trillion total parameters and 95 billion active per token, routed across 512 experts. It uses a hybrid attention design (Gated DeltaNet and Gated Attention layers) across 92 layers, with a native 262K context window and thinking mode enabled for every response.
Alibaba reports 93.0 on PaperBench (ahead of GPT-5.6 Sol's 90.5), 92.6 on GPQA Diamond, 86.6 on Terminal-Bench 2.1, and 67.7 on SWE-bench Pro, positioning it for coding, research, and long-horizon agentic work.
It gives developers access to Qwen-Max-class capability under open weights, useful for teams that want frontier-level coding and agentic performance without a closed API.
Context Window 1M
tokens
Max Output 131K
tokens
Input Cost $2
per million tokens
Output Cost $6
per million tokens
Input text
modalities
Tool Use Yes
Release Date Aug 12, 2026
Output Speed 39
tokens / sec
Latency 2.85s
time to first token
Try Qwen3.8 2.4T A95B for free
Try Qwen3.8 2.4T A95B instantly in your browser.
This playground uses the Puter.js AI API — no API keys or setup required.
Benchmarks
How Qwen3.8 2.4T A95B performs on standard evaluations.
| Benchmark | Score |
|---|---|
| GPQA Diamond Graduate-level science Q&A | 93.5% |
| Humanity's Last Exam Cross-domain reasoning | 42.4% |
| SciCode Scientific programming | 54.1% |
| LCR Long-context reasoning | 80.3% |
Scores sourced from Artificial Analysis.
Find other Qwen models →
Qwen3.8 Max Prime
Qwen3.8 Max Prime is a high-throughput serving tier of Alibaba's Qwen3.8 Max, the Qwen team's 2.4-trillion-parameter Mixture-of-Experts flagship released August 3, 2026. It runs the same underlying weights as standard Qwen3.8 Max, a 1M-token-context model with 95B active parameters and native text, image, and video input, but on faster inference infrastructure. Third-party trackers report roughly 1.5 to 2 times the output throughput of the standard endpoint, at a higher per-token price. Because the weights are unchanged, Prime inherits Qwen3.8 Max's benchmark results, including 67.7 on SWE-bench Pro and a top-three ranking on LMArena Code, putting it behind Claude Fable 5 but ahead of GPT-6 Sol on several agentic coding evaluations. It suits developers running latency-sensitive coding agents or high-concurrency workloads who want Qwen3.8 Max's capability without the standard endpoint's slower response time.
ChatQwen3.8 OmniFlash
Qwen3.8 OmniFlash is Alibaba's first omni-modal model built around agentic capabilities, released September 18, 2026, succeeding Qwen3.5-Omni-Plus. It accepts text, images, audio, and video as input and returns text, pairing native audio-video understanding with reasoning and tool use. It can watch or listen to content, plan a task, call tools, and deliver a finished result, such as an edited video or a meeting summary. Alibaba reports more than a 26% average improvement across 30 evaluations versus Qwen3.5-Omni-Plus, with gains in audio-video agents, coding, long-context tasks, and real-time interaction. It says audio-visual performance approaches Gemini 3.8 Flash, with audio performance exceeding it. With a 1,000,000 token context window, 128,000 token output limit, and pricing of $0.06 per million input tokens and $0.19 per million output tokens, it suits developers building video and audio agents, such as meeting summarization or video-editing pipelines that need tool-calling built in.
ChatQwen3.8 Max (0902)
Qwen3.8 Max (0902) is a September 2026 snapshot of Alibaba's Qwen3.8 Max, a 2.4-trillion-parameter mixture-of-experts flagship with roughly 95 billion active parameters per token. It keeps the 1M-token context window, thinking mode, and tool ecosystem of the base model, with additional post-training on coding and office-work tasks. On Code Arena WebDev it ranks first overall at 1,691 points, ahead of Claude Opus 5 Max (1,687) and Kimi K3 Max (1,674). The largest gains are on TerminalBench 3.0 and ProgramBench Almost Solved, both more than doubling versus the prior snapshot. Reported core scores include GPQA Diamond 92.6 and PaperBench 93.0. Against Claude Opus 5, it leads on several coding benchmarks and on WorkArena, though Opus 5 still leads on most agentic coding rows and both office-work benchmarks. Pricing is unchanged from the base Qwen3.8 Max, making it a straightforward upgrade for coding and office-automation workloads already on the platform.
Frequently Asked Questions
You can access Qwen3.8 2.4T A95B by Qwen through Puter.js AI API. Include the library in your web app or Node.js project and start making calls with just a few lines of JavaScript — no backend and no configuration required. You can also use it with Python or cURL via Puter's OpenAI-compatible API.
Qwen3.8 2.4T A95B is free to try with a Puter account. Every account includes a free AI allowance, and you can chat with it in the playground on this page. You can upgrade your account anytime for a larger allowance.
Qwen3.8 2.4T A95B is free to integrate using the Puter.js AI API. With the User-Pays Model, you can add AI to your app for $0, since users cover their own AI usage through their Puter account.
| Price per 1M tokens | |
|---|---|
| Input | $2 |
| Output | $6 |
Qwen3.8 2.4T A95B was created by Qwen and released on Aug 12, 2026.
Qwen3.8 2.4T A95B supports a context window of 1M tokens. For reference, that is roughly equivalent to 2,000 pages of text.
Qwen3.8 2.4T A95B can generate up to 131K tokens in a single response.
Qwen3.8 2.4T A95B accepts the following input types: text. It produces: text.
Yes, Qwen3.8 2.4T A95B supports tool use (function calling), allowing it to interact with external tools, APIs, and data sources as part of its response flow.
Yes — the Qwen3.8 2.4T A95B API works with any JavaScript framework, Node.js, or plain HTML through Puter.js. Just include the library and start building. See the documentation for more details.
Add Qwen3.8 2.4T A95B to your app for free
Developers can integrate Qwen3.8 2.4T A95B for free using the Puter.js AI API.
With the User-Pays Model, each user covers their own AI usage instead of the developer.