Qwen3.8 27B

obsidian/Qwen3.8-27B
New
VisionToolsJSONReasoning
by obsidian · 2026-08-15

Qwen3.8 27B served in block-FP8 for higher throughput and lower memory footprint, with the vision tower kept at full precision. Preserves the original model's reasoning, coding, multilingual performance, and tool use. Designed to provide direct, complete responses across a wide range of prompts. The model may occasionally append brief informational disclaimers inherited from the base model's training. Ideal for AI research, security testing, red teaming, agent development, coding assistants, and other advanced AI applications that benefit from maximum output flexibility.

Best for:reasoningvision
Endpoints:/v1/chat/completions/v1/responses
ctx262K tokens
Inputtext + image + video
Outputtext
Best forreasoning, vision
p50 TTFT3.95 s
INPUT$0.40/ 1M tokens
OUTPUT$4.21/ 1M tokens
p50 TTFT3.95 s7d
p95 TTFT10.00 s7d
TRAFFIC15842.9Mtokens / 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="obsidian/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$0.400
Output / 1M tokens$4.21
CurrencyUSD

Cost calculator

Tokens / month10MM
Input share70%%
Estimated / month $15.43

Estimate based on list price

Token & cost estimator

Input tokens: 20Cost per request: $0.002113

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

Performance

last 7 days
p50 TTFT
3.95 s
Output speed
42.9 tok/s
p95 TTFT
10.00 s
Error rate
7.9%

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
GPQA Diamond
90.5
Humanity's Last Exam
33.9
Long-Context Recall
77.3
SciCode
44.7
tau_banking
48.0
terminalbench_v2_1
79.8
Source: artificialanalysis.ai

Community buzz

What developers are saying this week

Hacker News11 mentions · 7ddown 5 vs the previous week

How it compares

Qwen3.8 27BQwen3.6 35B A3B Uncensored (Aggressive)Gemma4 26B A4B Uncensored (Balanced)
Input $/M$0.40$0.31$0.25
Output $/M$4.21$4.21$2.90
Context262K262K262K
Quality4/104/104/10
Compare side-by-sideCompare side-by-sideCompare side-by-side

Browse AI model comparisons →

FAQ

How much does Qwen3.8 27B cost on OrcaRouter?
Qwen3.8 27B is priced at $0.40 per 1M input tokens and $4.21 per 1M output tokens via OrcaRouter. Pricing is pulled live from the routing layer.
What is Qwen3.8 27B's context window?
Qwen3.8 27B supports a context window of 262K tokens. Use long-context features (RAG, summarisation) up to that limit.
How do I call 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="obsidian/Qwen3.8-27B" in the chat.completions.create call.
Does OrcaRouter rate-limit 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
Qwen3.8 27B•$0.40/M in•3946ms p50•via OrcaRouter
HTML <a href="https://www.orcarouter.ai/models/obsidian/Qwen3.8-27B" target="_blank"> <img src="https://www.orcarouter.ai/embed/obsidian/Qwen3.8-27B.svg" alt="Qwen3.8 27B on OrcaRouter" /> </a>
Markdown [![Qwen3.8 27B](https://www.orcarouter.ai/embed/obsidian/Qwen3.8-27B.svg)](https://www.orcarouter.ai/models/obsidian/Qwen3.8-27B)

Model card as data

GET /api/public/models/obsidian/Qwen3.8-27BOpen
Machine-readable:/llms.txt/llms-full.txt