Advertise here
#104

Qwen3.8 2.4T A95B

Qwen Release: 2026-08-12 Tested on: 2026-10-01 04:08 qwen/qwen3.8-2.4t-a95b::low
2.4T total (95B active)MoEWeights available
(high) (low)

Summary

Qwen3.8 2.4T A95B scores 8.4 on AI BENCHY and ranks #104. It has 9.6 reliability, a 73.9% pass rate, $4.016 total cost, and 120.40s average response time.

Model facts

Researched on 2026-08-12

Reported
Parameters
2.4T total (95B active)
Architecture
MoE
Availability
Weights available
License
Apache-2.0

Exact open-weight sparse checkpoint; public weights use the Qwen Apache-2.0 release license.

Parameter source Availability source

Score

8.4

Consistency

9.2

Reliability

9.6

Total Cost (Current Price)

$4.016 ↑ +26.6%

Tested at: $3.172

Total Output Tokens

424,019

Total Input Tokens

313,881

Input Price

$2.000 / 1M

Output Price

$6.000 / 1M

Cache Read Price

$0.250 / 1M

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: 7

Attempt pass rate: 73.9%

Flaky tests

2

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

Response Time (avg)

120.40s

Response Time (max): 534.21s

Response Time (total): 2769.16s

Wrong answer: 5 Invalid tool call: 1 No answer: 1

Hamster playing table tennis

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

#104 Qwen3.8 2.4T A95B

low
Cost
$0.100
Time
193.2s
Tokens
16,767 tok

Run history

Tested on Score Reliability Tests Correct Total Cost Compare
2026-10-01 04:08 Re-test 8.4 9.6 $4.016 ↑ Current run
2026-08-12 23:44 Initial run 8.2 9.1 $3.113 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 04:08 · Current run69/69 attempts8.49.29.616/232424,019313,881$4.016120.40s
2026-08-12 23:44 · Initial run66/66 attempts8.28.99.116/223554,485113,340$3.113132.39s
Difference—+0.3+0.3+0.50-1-130466+200541+$0.903-11990ms

Benchmark coverage differs: 69/69 attempts (Target: 3 repeats per test) versus 66/66 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.

Price History

Historical pricing data for this model from OpenRouter.

Date Input Price Output Price Cache Read Price Cache Write Price
2026-08-14 12:51 $2.000 / 1M $6.000 / 1M N/A N/A
2026-10-01 18:33 $2.000 / 1M $6.000 / 1M $0.250 / 1M N/A

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 2.4T A95BlowvsNemotron 3.5 Super VLhighFree AvailableQwen3.8 2.4T A95BlowvsGPT-5 MinimediumQwen3.8 2.4T A95BlowvsClaude Haiku 5.5maxQwen3.8 2.4T A95BlowvsMuse Glimmer 30BxhighQwen3.8 2.4T A95BlowvsQwen3.7 MaxmediumQwen3.8 2.4T A95BlowvsGemini 3.1 Pro PreviewmediumQwen3.8 2.4T A95BlowvsClaude Sonnet 5mediumQwen3.8 2.4T A95BlowvsInklinghighFree AvailableQwen3.8 2.4T A95BlowvsMuse Spark 1.3lowQwen3.8 2.4T A95BlowvsMuse Spark 1.1high

Category Breakdown

Category Score Consistency Tests Correct
Agentic 10.0 10.0
Anti-AI Tricks 10.0 10.0
Coding 7.8 9.3
Combined 8.2 6.9
Data parsing and extraction 10.0 10.0
Domain specific 3.2 9.3
General Intelligence 6.1 3.1
Instructions following 9.8 10.0
Puzzle Solving 10.0 10.0
Tool Calling 10.0 10.0
Trivia 3.0 10.0

Compared models