Qwen3.8 27B

Qwen3.8 27B is an open-weight multimodal model from Alibaba for coding, visual understanding, tool use, and structured output. This variant keeps thinking disabled for faster direct responses. Requests containing video cost $0.32 per million input tokens and $2.50 per million output tokens.

Reasoning
Vision
Video Input
Tool Calling
Structured Output

Added Aug 16, 2026

Model weights

Pricing

Auto routing · per 1M tokens

Input: $0.15
Output: $0.70
Cache read: $0.040

Compare provider prices

Specifications

Context window: 1M
Max output: 32.8K
Parameters: 27B
Avg output (7d): 325 tokens; Longer than 27% of models

Benchmarks

Artificial Analysis

LMArena

Sourced from Artificial Analysis.

Intelligence Index

20.2

Better than 68% of models compared

Coding Index

44.6

Better than 51% of models compared

Agentic Index

22.4

Better than 61% of models compared

Agentic work

AutomationBench-AA

Workflow automation with guardrail penalties

12.4%

Better than 36% of models compared

AA-Briefcase

Agentic knowledge work (Elo)

898 Elo

Better than 41% of models compared

GDPval-AA v2

Economically valuable tasks (Elo)

1074 Elo

Better than 51% of models compared

Document reasoning

GDP.pdf

Professional PDF reasoning: all-pass rate

5.2%

Better than 25% of models compared

AA-LCR v1.1

Long context reasoning with updated grading

69.3%

Better than 61% of models compared

Reasoning

HLE

Humanity's Last Exam

12.1%

Better than 56% of models compared

CritPt

Research-level physics reasoning

0.3%

Coding

Terminal-Bench v4.0

Practical coding and terminal tasks

0.0%

Better than 16% of models compared

SciCode

Python programming for scientific computing

36.2%

Better than 15% of models compared

Knowledge

AA-Omniscience Accuracy

Proportion of correctly answered questions

8.6%

AA-Omniscience Hallucination Rate

Rate of incorrect answers among non-correct responses

18.1%

Legacy benchmarks

GPQA Diamond (legacy)

Graduate-level scientific reasoning

81.8%

Better than 70% of models compared

AA-LCR (unversioned / legacy)

Long context reasoning evaluation

69.3%

Better than 60% of models compared

GDPval-AA (unversioned / legacy)

Economically valuable tasks

28.6%

Last updated Oct 14, 2026