Qwen3.8 27B

qwen/qwen3.8-27b
NewFeatured
VisionToolsJSONReasoning
by Qwen · 2026-08-13

Qwen3.8-27B is Alibaba's open-weight 27B dense multimodal model, released under Apache-2.0 and self-hosted on OrcaRouter's own infrastructure. It accepts text, images, and video and returns text, with a 256K-token context window (262,144 positions) built on the Qwen3.5 architecture - 64 layers, 5120 hidden size, with a dedicated vision tower. Despite its 27B size it posts unusually strong agentic and coding results on Qwen's own model card, including 90.3 on LiveCodeBench v6, 89.2 on GPQA Diamond, 84.3 on OSWorld-Verified, and 79.0 on QwenSWEBench - competitive with far larger closed models on several axes. It supports native tool calling, structured outputs, reasoning mode, and the full sampling surface. Because the weights are open and we run them ourselves, there is no per-token vendor cost to pass through. That makes Qwen3.8-27B a strong default for high-volume multimodal work: document and screenshot understanding, video frames, coding agents, and computer-use style automation.

Endpoints:/v1/chat/completions/v1/responses
ctx262K tokens
Inputtext + image + video
Outputtext
p50 TTFT1.75 s
INPUT$0.33/ 1M tokens
OUTPUT$2.40/ 1M tokens
p50 TTFT1.75 s7d
p95 TTFT10.00 s7d
TRAFFIC1166.5Mtokens / 7d
Get the Qwen3.8 27B API →▶ Try in playground</> Use via API

Code samples

Call from any SDK

OpenAI-compatible — keep the SDK you already use

  • OpenAI SDKhttps://api.orcarouter.ai/v1
import os

from openai import OpenAI

client = OpenAI(
    base_url="https://api.orcarouter.ai/v1",
    api_key=os.environ["ORCAROUTER_API_KEY"],
)

response = client.chat.completions.create(
    model="qwen/qwen3.8-27b",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Supported parameters

  • frequency_penalty
  • include_reasoning
  • logprobs
  • max_tokens
  • min_p
  • presence_penalty
  • reasoning
  • repetition_penalty
  • response_format
  • seed
  • stop
  • structured_outputs
  • temperature
  • tool_choice
  • tools
  • top_k
  • top_p

Pricing

Input / 1M tokens$0.330
Output / 1M tokens$2.40
CurrencyUSD

Cost calculator

Tokens / month10MM
Input share70%%
Estimated / month $9.51

Estimate based on list price

Token & cost estimator

Input tokens: 20Cost per request: $0.001207

Estimate only — actual token counts depend on the provider's tokenizer.

Performance

last 7 days
p50 TTFT
1.75 s
Output speed
81.5 tok/s
p95 TTFT
10.00 s
Error rate
12.0%

Public benchmarks

Last evaluated 2026-08-14
68.1
AA Coding
Better than 77% of models compared
#30 of 133
52.0
AA Intelligence
Better than 77% of models compared
#29 of 135
AndroidWorld
81.9
CharXiv (RQ)
78.8
CoWorkBench
70.7
DeepSWE 1.1
42.2
ERQA
65.5
GPQA Diamond
90.5
HLE
30.8
Humanity's Last Exam
33.9
IFBench
79.5
JobBench
33.4
LiveCodeBench v6
90.3
Long-Context Recall
77.3
NL2Repo-Bench
42.3
OmniDocBench 1.5
91.1
OSWorld-Verified
84.3
QwenSWEBench
79.0
RealWorldQA
85.9
RecreationBench
47.1
SciCode
44.7
SWE-bench Pro
61.7
SWE-MM
38.6
tau_banking
48.0
Terminal Bench 2.1 (Terminus)
73.0
terminalbench_v2_1
79.8
Vision2Web
62.9
WebArena-Verified
64.8
Source: artificialanalysis.ai, huggingface.co/Qwen

Community buzz

What developers are saying this week

Hacker News11 mentions · 7ddown 5 vs the previous week

How it compares

Qwen3.8 27BQwen3.8 Maxqwen/qwen3-max-previewQwen3.5 397B A17B
Input $/M$0.33$2.00$0.86$0.17
Output $/M$2.40$6.00$3.44$1.03
Context262K1.0M262K33K
Quality4/109/108/108/10
Compare side-by-sideCompare side-by-sideCompare side-by-sideCompare side-by-side

Browse AI model comparisons →

More from Qwen

See all models from qwen →
qwen/qwen3.5-plusCheapest
qwen/qwen3.5-plus
$0.12 in · $0.69 out / 1M
1M ctx· quality 8/10
 Compare side-by-side
Qwen3.6 35B A3B
qwen/qwen3.6-35b-a3b
$0.25 in · $1.49 out / 1M
262K ctx· quality 8/10
 Compare side-by-side
Qwen3.6 Plus
qwen/qwen3.6-plus
$0.50 in · $3.00 out / 1M
1M ctx· quality 8/10
 Compare side-by-side

FAQ

How much does Qwen: Qwen3.8 27B cost on OrcaRouter?
Qwen: Qwen3.8 27B is priced at $0.33 per 1M input tokens and $2.40 per 1M output tokens via OrcaRouter. Pricing is pulled live from the routing layer.
What is Qwen: Qwen3.8 27B's context window?
Qwen: Qwen3.8 27B supports a context window of 262K tokens. Use long-context features (RAG, summarisation) up to that limit.
How do I call Qwen: Qwen3.8 27B via the OpenAI SDK?
Set OpenAI base_url to https://api.orcarouter.ai/v1, supply your OrcaRouter API key, and pass model="qwen/qwen3.8-27b" in the chat.completions.create call.
Does OrcaRouter rate-limit Qwen: Qwen3.8 27B?
Per-model rate limits follow your OrcaRouter plan. Free tiers ship with conservative caps; paid tiers lift them. Check /pricing for current quotas.

Embed this badge

Paste into your blog post
Qwen: Qwen3.8 27B•$0.33/M in•1755ms p50•via OrcaRouter
HTML <a href="https://www.orcarouter.ai/models/qwen/qwen3.8-27b" target="_blank"> <img src="https://www.orcarouter.ai/embed/qwen/qwen3.8-27b.svg" alt="Qwen: Qwen3.8 27B on OrcaRouter" /> </a>
Markdown [![Qwen: Qwen3.8 27B](https://www.orcarouter.ai/embed/qwen/qwen3.8-27b.svg)](https://www.orcarouter.ai/models/qwen/qwen3.8-27b)

Model card as data

GET /api/public/models/qwen/qwen3.8-27bOpen
Machine-readable:/llms.txt/llms-full.txt