The LLM benchmark aggregator.

Qwen3.8 Flash

Alibaba · 180B · open weights

Strongest in Web research (#21 of 48, on 1 of its 2 boards), weakest in Agents (#124 of 289). Above par in 8 of 13 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5 and GPT-6 Astra and ahead of DeepSeek V4 Flash Vision and Step 5.

§ 1 · Profile

What it is good at

Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.

Coding52.8−13.6#59/1761/6

Agentic coding53.1−13.9#55/1691/5

Reasoning51.8−13#81/1891/5

Reasoning52.9−13.8#51/1431/3

Core abilities51.2−14#87/2251/4

Data analysis52.2−13.7#28/471/1

General intelligence50.7−18.1#94/2141/3

Professional49.6−13.5#99/1871/2

Document design48.5−18.6#31/471/1

Agents49.9−16#124/2891/6

Web research52.9−10.8#21/481/2

Long-horizon work48.5−25.5#26/471/1

Computer use48.2−19.7#44/801/2

§ 2 · Head to head

What it beats, and what beats it

The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Qwen3.8 Flash, left for the other.

GPT-5.4 Nanowon 8 · lost 0

GPT-4o Miniwon 8 · lost 0

GPT-4.1 Nanowon 8 · lost 0

Llama 4 Maverick Instructwon 8 · lost 0

GPT OSS 120Bwon 8 · lost 0

DeepSeek V3won 8 · lost 0

NVIDIA Nemotron 3.5 Lightning 30B A3Bwon 13 · lost 0

Claude Haiku 4.5won 13 · lost 0

Gemma 4 31Bwon 13 · lost 0

NVIDIA Nemotron 3 Ultra 550B A55Bwon 13 · lost 0

Nemotron 3.5 Super VLwon 13 · lost 0

Beam 1.0won 13 · lost 0

Inklingwon 12 · lost 1

GPT-5.4 Miniwon 7 · lost 1

Gemini 3.5 Flash Litewon 7 · lost 1

GPT-4.1 Miniwon 7 · lost 1

Mistral Large 3won 7 · lost 1

DeepSeek V3.2won 7 · lost 1

Mistral Large 4won 11 · lost 2

MiniMax M3won 10 · lost 3

Muse Spark Openwon 10 · lost 3

Inkling Smallwon 6 · lost 2

GPT-4.1won 6 · lost 2

Kimi K2won 5 · lost 3

Gemini 3.1 Prowon 8 · lost 5

Claude Sonnet 4won 5 · lost 4

Claude Sonnet 4.6won 7 · lost 6

GLM-5.3 Flashwon 7 · lost 6

Step 5won 7 · lost 6

DeepSeek V4 Flash Visionwon 7 · lost 6

DeepSeek R1won 4 · lost 4

GPT-5 Miniwon 4 · lost 4

Gemini 2.5 Prowon 4 · lost 4

GPT-5.6 Lunawon 5 · lost 8

GPT-6 Lunawon 5 · lost 8

Qwen3.8 27Bwon 5 · lost 8

Muse Spark 1.2won 5 · lost 8

Gemini 3.7 Flashwon 5 · lost 8

Gemini 2.5 Flashwon 3 · lost 5

GPT-5.2won 3 · lost 5

GPT-5won 3 · lost 5

Grok 4won 3 · lost 5

Gemini 3.5 Flashwon 3 · lost 5

Claude Sonnet 4.5won 3 · lost 6

O3won 3 · lost 6

Kimi K2.5won 3 · lost 6

GLM-5.3won 4 · lost 9

Qwen3.8 Maxwon 4 · lost 9

Grok 4.5won 4 · lost 9

GLM-5.2won 2 · lost 6

DeepSeek V4 Flashwon 2 · lost 6

Claude Opus 4.5won 2 · lost 6

O4 Miniwon 2 · lost 6

Gemini 3 Prowon 2 · lost 6

Claude Haiku 5.5won 3 · lost 10

Gemini 3.8 Flashwon 3 · lost 10

Mimo V2.6 Prowon 3 · lost 10

Claude Sonnet 5won 2 · lost 11

Grok 4.7won 2 · lost 11

GPT-5.5won 2 · lost 11

DeepSeek V4 Prowon 1 · lost 7

GPT-5.6 Terrawon 1 · lost 12

Claude Sonnet 5.5won 1 · lost 12

Claude Opus 4.8won 1 · lost 12

Kimi K3won 1 · lost 12

GPT-6 Solwon 1 · lost 12

GPT-5.6 Solwon 1 · lost 12

Muse Spark 1.3won 1 · lost 12

Claude Opus 4.7won 0 · lost 8

GPT-5.4won 0 · lost 8

Claude Opus 5won 0 · lost 13

Grok 4.6won 0 · lost 13

DeepSeek V4.1 Flashwon 0 · lost 13

Claude Opus 5.5won 0 · lost 13

Muse Wonderwon 0 · lost 13

Claude Fable 5.1won 0 · lost 13

Gemini 4 Argonwon 0 · lost 13

GPT-6.1 Solwon 0 · lost 13

GPT-6 Astrawon 0 · lost 13

Claude Fable 5won 0 · lost 13

§ 3 · Sources

Where the numbers come from

1 publication, 8 figures. Every one links to the page it was read from.

LiveBench ↗8 measuresread 2026-10-19

LiveBench 68.8LiveBench · Agentic 37.1LiveBench · Reasoning 87.4LiveBench · Coding 87.4LiveBench · Browser Use 48LiveBench · Data Analysis 74.2LiveBench · Research 74.6LiveBench · Design 72.8

Badge

[![PublicAI Index](https://publicai.io/model-index/badge?model=qwen3-8-flash)](https://publicai.io/model-index/m/qwen3-8-flash)