Qwen3.8-27B
Compact 27B multimodal reasoning model with a 262K context window, suited to coding, tool use, long-context work, and vision tasks.
Parameters
27B
Dense · FP8
Context Window
262K
Tokens
Intelligence
52
AA Index
Released
Aug 2026
13 Aug 2026
A small model that reasons like a big one
Qwen3.8-27B is a compact multimodal reasoning model from Alibaba's Qwen family, designed for general-purpose reasoning, coding, tool use, and vision workloads. Its 262K-token context window supports long documents and extended agentic tasks, while the FP8 deployment offers efficient serving.
Thinking mode: this model reasons step-by-step before responding by default. Set reasoning_effort to none to disable thinking. Other effort levels enable thinking but do not select graduated reasoning budgets.
Built for high-volume async workloads
Complex Reasoning
Reasons step-by-step before responding by default, producing structured answers on hard analytical tasks.
Code Generation
Strong coding and tool-use behaviour at a compact 27B footprint, ideal for high-volume async pipelines.
Tool Use & Agents
Reliable function calling for extended agentic runs where every token counts.
Vision Tasks
Multimodal input handling for document understanding, screenshots, and image-grounded reasoning.
Long-Context Workloads
262K tokens covers full repositories, long legal documents, and multi-document synthesis.
Efficient FP8 Serving
The FP8 deployment keeps throughput high and per-token cost low for batch and async jobs.
Flexible pricing tiers
Choose the optimal balance of speed and cost for your workflow. Prices are per 1M tokens.
| Tier | Input / 1M tokens | Output / 1M tokens |
|---|---|---|
| Batch (24h)Cheapest | $0.25 | $1.50 |
| Async | $0.35 | $2.25 |
| Realtime | $0.45 | $3.00 |
Context window natively supported up to 262K tokens. FP8 quantization.
We can offer custom pricing for bulk discounts, large workloads, and dedicated deployments - reach out to hello@doubleword.ai.
Start building in minutes
Qwen3.8-27B is accessible via OpenAI-compatible endpoints. Here's how to integrate it with the standard Python SDK.
from openai import OpenAI
client = OpenAI(
api_key="your-api-key-here",
base_url="https://api.doubleword.ai/v1"
)
# Thinking is enabled by default
response = client.chat.completions.create(
model="Qwen/Qwen3.8-27B-FP8",
messages=[
{"role": "user", "content": "Summarise this contract's termination clauses."}
],
# Disable step-by-step reasoning:
# reasoning_effort="none",
)
print(response.choices[0].message.content)💡 Pro tip
Thinking is on by default. For simple, latency-sensitive calls set reasoning_effort="none" — other effort levels keep thinking on but do not set graduated reasoning budgets.