---
id: "qwen/qwen3.8-2.4t-a95b"
also: ["qwen/qwen3.8-2.4t-a95b-20260812", "Qwen/Qwen3.8-2.4T-A95B", "hf/Qwen/Qwen3.8-2.4T-A95B", "requesty/alibaba/qwen3.8-2.4t-a95b", "requesty/deepinfra/qwen3.8-2.4T-A95B", "vercel/alibaba/qwen3.8-2.4t-a95b"]
cheapest_list_usd_per_m: "2.0 in, 6.0 out, at deepinfra on hf"
cheapest_effective_usd_per_m: "0.3546 in, 6.0 out, at together on openrouter"
context: 1048576
max_output: 131072
input: ["text"]
output: ["text"]
released: "2026-08-12"
reasoning: ["xhigh", "medium", "low"]
api: ["/api/v1/chat/completions", "/api/v1/completions", "/api/v1/messages", "/api/v1/responses"]
params: ["frequency_penalty", "include_reasoning", "logit_bias", "logprobs", "max_tokens", "min_p", "presence_penalty", "reasoning", "reasoning_effort", "repetition_penalty", "response_format", "seed", "stop", "structured_outputs", "temperature", "tool_choice", "tools", "top_k", "top_logprobs", "top_p"]
endpoints: 27
gateways: "openrouter (10), vercel (10), hf (4), requesty (3)"
no_training: 17
trains: 3
retains_prompts: 5
coding: 71.9000
agentic: 50.1000
---
# Qwen: Qwen3.8 2.4T A95B

Qwen 3.8 Max is a 2.4-trillion-parameter MoE model delivering a comprehensive leap in coding and professional work. Autonomously codes and delivers complete projects spanning 10+ days. Handles hundreds of specialized tasks across legal, financial, design, and other professional domains, producing production-grade results end-to-end in a single conversation. Native visual understanding runs through the full cycle of planning, execution, and verification, enabling deep semantic analysis of ultra-long documents and extended video content. In long-horizon tasks, plans autonomously, iterates through closed feedback loops, and continuously evolves.