Advertise here
#402

Qwen3.5-9B

Qwen Release: 2026-03-02 Tested on: 2026-10-01 01:14 qwen/qwen3.5-9b::medium
9BDenseOpen source
(medium) (none)

Summary

Qwen3.5-9B scores 4.0 on AI BENCHY and ranks #402. It has 8.7 reliability, a 26.1% pass rate, $0.106 total cost, and 110.18s average response time.

What makes Qwen3.5-9B unique: It stands out most in Agentic, where it ranks #2, while General Intelligence is its weakest area at #10. It uses unusually many reasoning tokens, which can help explain its slower or more expensive runs.

Model facts

Researched on 2026-08-12

Reported
Parameters
9B
Architecture
Dense
Availability
Open source
License
Apache-2.0
Parameter source Availability source

Score

4.0

Consistency

7.9

Reliability

8.7

Total Cost (Current Price)

$0.106

Total Output Tokens

447,247

Total Input Tokens

387,323

Input Price

$0.100 / 1M

Output Price

$0.150 / 1M

Cache Read Price

N/A

Cache Write Price

N/A

Cache prices apply to input tokens. Reads reuse cached prompts; writes store them and can cost extra. Output tokens use the output price.

Tests Correct

Wrong Tests: 20

Attempt pass rate: 26.1%

Flaky tests

6

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

Response Time (avg)

110.18s

Response Time (max): 569.97s

Response Time (total): 1873.00s

Timed out: 11 No answer: 3 Wrong answer: 2 API error: 1 Extra formatting: 1 Did not follow instructions: 1 Invalid tool call: 1

Hamster playing table tennis

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

#402 Qwen3.5-9B

medium
Cost
$0.001
Time
35.9s
Tokens
3,030 tok

Run history

Tested on Score Reliability Tests Correct Total Cost Compare
2026-10-01 01:14 Re-test 4.0 8.7 $0.106 Current run
2026-07-16 23:12 New test added 3.8 5.0 $0.036 ↑ Compare
2026-06-04 13:41 New test added 4.2 5.6 $0.035 ↓ Compare
2026-05-22 00:18 Suite changed 4.2 1.7 $0.035 Compare
2026-05-08 14:44 Suite changed 4.3 3.3 $0.035 Compare
2026-05-08 14:44 Suite changed 4.3 3.3 $0.035 Compare
2026-04-20 17:48 First recorded run 4.4 N/A $0.030 Compare

This run used a different benchmark suite. Keep suite changes in mind when reading historical movement.

Run comparison

RunBenchmark coverageScoreConsistencyReliabilityTests CorrectFlaky testsTotal Output TokensTotal Input TokensTotal CostResponse Time (avg)
2026-10-01 01:14 · Current run69/69 attempts4.07.98.73/236447,247387,323$0.106110.18s
2026-06-04 13:41 · New test added63/63 attempts4.28.05.63/215238,56117,070$0.03582.24s
Difference—-0.1-0.1+3.10+1+208686+370253+$0.071+27934ms

Benchmark coverage differs: 69/69 attempts (Target: 3 repeats per test) versus 63/63 attempts (Target: 3 repeats per test). Totals and repeat-sensitive metrics are not directly comparable.

These two runs used different benchmark suites, so the deltas reflect both model changes and suite changes.

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

Qwen3.5-9BmediumvsMercury 2.5noneQwen3.5-9BmediumvsLing 3.0 TinydefaultQwen3.5-9Bmediumvsgpt-oss-120bnoneQwen3.5-9BmediumvsGranite 4.1 8bdefaultQwen3.5-9BmediumvsGLM 4.7 FlashmediumQwen3.5-9BmediumvsNemotron 3.5 LightningnoneFree AvailableQwen3.5-9BmediumvsMiniMax M2.5mediumQwen3.5-9BmediumvsXing4.0 29b A4bmediumFree AvailableQwen3.5-9BmediumvsMercury 2noneQwen3.5-9BmediumvsLing 3.0 Tinyhigh

Category Breakdown

Category Score Consistency Tests Correct
Agentic 5.2 3.2
Anti-AI Tricks 5.1 5.8
Coding 2.9 10.0
Combined 3.0 10.0
Data parsing and extraction 3.6 5.6
Domain specific 3.6 7.2
General Intelligence 2.8 1.6
Instructions following 6.5 10.0
Puzzle Solving 3.0 10.0
Tool Calling 10.0 10.0
Trivia 3.0 10.0

Compared models