AI BENCHY
Advertise here
#42

Qwen3.8 27B

Qwen Release: 2026-08-14 Tested on: 2026-08-15 04:08 qwen/qwen3.8-27b::high
27.3BOtherWeights available
(high) (medium) (low) (none)

Summary

Qwen3.8 27B scores 8.4 on AI BENCHY and ranks #42. It has 9.6 reliability, a 77.3% pass rate, N/A total cost, and 167.89s average response time.

What makes Qwen3.8 27B unique: It uses unusually many reasoning tokens, which can help explain its slower or more expensive runs.

Model facts

Researched on 2026-08-14

Reported
Parameters
27.3B
Architecture
Other
Availability
Weights available
License
Apache-2.0

Exact local Q4_K_M GGUF reports 27.3B parameters and a hybrid attention/SSM architecture through Ollama model details. Tested locally on an NVIDIA GeForce RTX 3090.

Parameter source Availability source

Score

8.4

Consistency

9.2

Reliability

9.6

Total Cost (Current Price)

N/A

Total Output Tokens

656,635

Total Input Tokens

100,179

Input Price

N/A

Output Price

N/A

Tests Correct

Wrong Tests: 6

Attempt pass rate: 77.3%

Flaky tests

2

Flaky tests had mixed outcomes across runs (at least one pass and one fail).

Response Time (avg)

167.89s

Response Time (max): 640.85s

Response Time (total): 3693.54s

Did not follow instructions: 2 No answer: 2 Wrong answer: 2

Hamster playing table tennis

Prompt: Create a detailed SVG illustration of a hamster playing table tennis.

#42 Qwen3.8 27B

high
Cost
N/A
Time
270.1s
Tokens
18,963 tok

Charts

Choose the first model, then click a second model to open a side-by-side page.

Total Output Tokens

Score vs Total Output Tokens

Quick Compare

Qwen3.8 27BhighvsGPT-5.2mediumQwen3.8 27BhighvsSeed 2.1 TurbomediumQwen3.8 27BhighvsGrok 4.5lowQwen3.8 27BhighvsInkling SmallhighQwen3.8 27BhighvsGrok 4.6mediumQwen3.8 27BhighvsMuse Spark 1.1lowQwen3.8 27BhighvsMuse Spark 1.2lowQwen3.8 27BhighvsGemini 2.5 FlashmediumQwen3.8 27BhighvsGrok 4.5mediumQwen3.8 27BhighvsClaude Sonnet 5medium

Category Breakdown

Category Score Consistency Tests Correct
Anti-AI Tricks 10.0 10.0
Coding 8.1 7.0
Combined 10.0 10.0
Data parsing and extraction 10.0 10.0
Domain specific 5.3 10.0
General Intelligence 5.0 10.0
Instructions following 9.8 10.0
Puzzle Solving 7.6 7.2
Tool Calling 10.0 10.0
Trivia 3.0 10.0

Compared models