alibaba
Qwen3.8 27B
Built on the architectural foundation of Qwen3.5, Qwen3.8 delivers substantial gains across coding, professional work, research, and long-horizon agentic tasks. Qwen3.8-27B brings these advances to a compact, deployment-friendly dense model: a native vision-language model that understands images and videos, with flexible thinking control, designed to carry complex, multi-step tasks through to completion with greater reliability.
reasoningtool usevisionimplicit cachingvideo inputfile inputexplicit cachingstructured outputtextimagepdfvideo
Overview
- Type
- language
- Context
- 1M
- Max output
- 131K
- Input / output
- text, image, pdf, video / text
API
Model identifier
alibaba/qwen3.8-27bProviders and live performance
Rolling Gateway metrics. Latency is time to first token; throughput is output tokens per second.
| Provider | Status | Latency | Throughput | Uptime 15m / 1h / 1d | Input | Output | Cache read / write | Discount | Regions | ZDR | No training |
|---|---|---|---|---|---|---|---|---|---|---|---|
| alibaba | Operational | 2.67s / 7.92s | 63.5 / 69.7 tok/s | 100.00% / 100.00% / 99.99% | $0.50/M | $3.00/M | $0.10/M / $0.63/M | — | — | Yes | Yes |
| cerebras | Operational | 0.52s / 3.61s | 1077.0 / 1195.8 tok/s | 100.00% / 100.00% / 99.96% | $0.99/M | $1.49/M | $0.99/M / — | — | — | No | Yes |
| deepinfra | Operational | 0.64s / 1.42s | 57.0 / 64.0 tok/s | 100.00% / 100.00% / 99.86% | $0.15/M | $1.88/M | $0.04/M / — | — | — | Yes | Yes |
| novita | Operational | 2.28s / 6.62s | 54.5 / 64.8 tok/s | 100.00% / 100.00% / 99.91% | $0.42/M | $3.00/M | $0.08/M / — | — | — | — | No |
| parasail | Operational | 0.69s / 1.27s | 80.0 / 119.6 tok/s | 100.00% / 100.00% / 100.00% | $0.24/M | $2.20/M | $0.05/M / — | — | — | Yes | Yes |
| runinfra | Operational | 1.04s / 3.18s | 185.0 / 208.9 tok/s | 100.00% / 100.00% / 100.00% | $0.10/M | $0.40/M | $0.01/M / — | — | — | No | No |
| wafer | Operational | 0.57s / 1.17s | 117.0 / 117.0 tok/s | 100.00% / 100.00% / 100.00% | $0.20/M | $6.00/M | $0.20/M / — | — | — | Yes | Yes |
Pricing
input
0.0000005
output
0.000003
input cache read
0.0000001
input cache write
0.000000625
varies by provider
Yes
Capabilities
reasoningtool usevisionimplicit cachingvideo inputfile inputexplicit cachingstructured outputtextimagepdfvideo
Supported parameters
max tokens
Yes
temperature
Yes
stop
Yes
tools
Yes
tool choice
Yes
reasoning
Yes
include reasoning
Yes
response format
Yes
structured outputs
Yes
Reasoning options
options
{
"type": "effort",
"values": [
"none",
"low",
"medium",
"xhigh"
]
}
Metadata
type
language
context window
1M
max output
131K
released
Aug 14, 2026
temperature control
Yes
zero data retention
some
no training
some
api specifications
v2, v3, v4