Noometry

Alibaba (Qwen), open weights

Qwen3-Next 80B-A3B Instruct

Qwen3-Next 80B-A3B Instruct by Alibaba (Qwen) ranks 102nd of 354 ranked models on the Noometry Index as of October 2026, with a score of 43.0. Its strongest category is reasoning, where it ranks 81st. API pricing starts at $0.50 per million input tokens and $2 per million output tokens, with a 131K-token context window.

Last verified

Noometry rank
#102 of 354
Index score
43.0
Evidence
Confirmed 25 results
Provider
Alibaba (Qwen)
Released
September 1, 2025
Weights
Open weights
Reasoning
No
Context window
131K
Max output
33K
Input price
$0.50 / M
Output price
$2 / M
Blended price
$0.88 / M
Output speed
111 tokens/s Kagi
Value
#95 of 219
Knowledge cutoff
April 2025
Input
text
Hugging Face
Qwen/Qwen3-Next-80B-A3B-Instruct

Category scores

Each category score combines every public result we have in that category.

  1. Coding 42.5
  2. Reasoning 31.1
  3. Math 38.8
  4. Knowledge 41.8
  5. Multilingual 52.1
  6. Instruction Following 70.8
  7. Long Context 37.0
  8. Writing & Preference 58.0
CategoryScoreRankResults
Coding42.5#981
Reasoning31.1#812
Math38.8#1262
Knowledge41.8#1064
Multilingual52.1#931
Instruction Following70.8#1592
Long Context37.0#2232
Writing & Preference58.0#1214

Strengths and weaknesses

Categories where Qwen3-Next 80B-A3B Instruct places highest and lowest among the models ranked in each, with its score against that category's median.

Strongest categories

CategoryScorevs medianRank
Reasoning31.1+7.5#81 of 350, top 24%
Coding42.5+3.8#98 of 340, top 29%
Multilingual52.1+4.7#93 of 297, top 32%

Weakest categories

CategoryScorevs medianRank
Long Context37.0−3.9#223 of 296, top 76%
Instruction Following70.8−0.5#159 of 305, top 53%
Writing & Preference58.0+4.3#121 of 312, top 39%

Closest competitors

The models ranked just above and below Qwen3-Next 80B-A3B Instruct. When scores are this close, price and speed are often the better way to choose.

ModelRankScoreBlended $/MSpeed
Hunyuan Vision 1.5#9843.1——Compare
Mistral Large 4#9943.1$1.03—Compare
Claude Opus 4#10043.1$3029Compare
Amazon Nova Experimental Chat 11 10#10143.0——Compare
MiMo-V2-Pro#10343.0$0.54—Compare
Amazon Nova Experimental Chat 12 10#10442.9——Compare
o3-pro#10542.9$351Compare
Qwen3.5 Plus#10642.9$0.90—Compare

Sponsored placements are available on pages like this one. Advertise on Noometry

Benchmark results

Every published result we track, with its source. Bold rows are the ones used for ranking; where several exist we prefer independent runs over self-reported numbers.

Coding

BenchmarkScorePositionSettingSourceDate
LMArena Coding1440#97 of 294, top 33%LMArena2026-10-08
LMArena Coding1391thinkingLMArena2026-10-08

Reasoning

BenchmarkScorePositionSettingSourceDate
Kagi LLM Benchmark66.7%#31 of 99, top 32%Kagi LLM Benchmark
Kagi LLM Benchmark54.4%Kagi LLM Benchmark
LMArena Hard Prompts1428#93 of 297, top 32%LMArena2026-10-08
LMArena Hard Prompts1371thinkingLMArena2026-10-08

Math

BenchmarkScorePositionSettingSourceDate
Omni-MATH46.7%#19 of 57, top 34%HELM Capabilities
LMArena Math1440#71 of 285, top 25%LMArena2026-10-08
LMArena Math1398thinkingLMArena2026-10-08

Knowledge

BenchmarkScorePositionSettingSourceDate
MMLU-Pro78.6%#20 of 58, top 35%HELM Capabilities
Vectara Hallucination Rate (lower is better)9.3%#48 of 96, top 50%Vectara Hallucination Leaderboard
GPQA (HELM)63%#20 of 57, top 36%HELM Capabilities
LMArena Expert1417#110 of 273, top 41%LMArena2026-10-08
LMArena Expert1376thinkingLMArena2026-10-08

Multilingual

BenchmarkScorePositionSettingSourceDate
LMArena Non-English1407#93 of 297, top 32%LMArena2026-10-08
LMArena Non-English1342thinkingLMArena2026-10-08
LMArena Chinese1460#92 of 285, top 33%LMArena2026-10-08
LMArena Chinese1406thinkingLMArena2026-10-08
LMArena French1413#107 of 223, top 48%LMArena2026-10-08
LMArena French1351thinkingLMArena2026-10-08
LMArena German1417#81 of 231, top 36%LMArena2026-10-08
LMArena German1356thinkingLMArena2026-10-08
LMArena Japanese1395#63 of 211, top 30%LMArena2026-10-08
LMArena Japanese1280thinkingLMArena2026-10-08
LMArena Korean1364#89 of 213, top 42%LMArena2026-10-08
LMArena Korean1310thinkingLMArena2026-10-08
LMArena Russian1404#102 of 283, top 37%LMArena2026-10-08
LMArena Russian1338thinkingLMArena2026-10-08
LMArena Spanish1435#73 of 226, top 33%LMArena2026-10-08
LMArena Spanish1358thinkingLMArena2026-10-08

Instruction Following

BenchmarkScorePositionSettingSourceDate
IFEval81%#40 of 57, top 71%HELM Capabilities
LMArena Instruction Following1389#113 of 298, top 38%LMArena2026-10-08
LMArena Instruction Following1344thinkingLMArena2026-10-08

Long Context

BenchmarkScorePositionSettingSourceDate
Fiction.LiveBench41.7%Epoch AI
Fiction.LiveBench55.6%#32 of 47, top 69%Epoch AI
LMArena Longer Query1403#114 of 291, top 40%LMArena2026-10-08
LMArena Longer Query1353thinkingLMArena2026-10-08

Writing & Preference

BenchmarkScorePositionSettingSourceDate
LMArena Text1417#101 of 297, top 35%LMArena2026-10-08
LMArena Text1368thinkingLMArena2026-10-08
LMArena Creative Writing1334#141 of 295, top 48%LMArena2026-10-08
LMArena Creative Writing1315thinkingLMArena2026-10-08
WildBench80.7%#26 of 57, top 46%HELM Capabilities
LMArena Multi-Turn1416#101 of 295, top 35%LMArena2026-10-08
LMArena Multi-Turn1346thinkingLMArena2026-10-08

API pricing by provider

RouteInput $/MOutput $/MCached input $/MChecked
alibaba$0.50$2—2026-10-10
bedrock$0.15$1.20—2026-10-10
deepinfra$0.09$1.10—2026-10-10
openrouter$0.10$1.10$0.072026-10-10

All Alibaba (Qwen) API prices → Estimate your cost →

Compare Qwen3-Next 80B-A3B Instruct

Other Alibaba (Qwen) models

Frequently asked questions

How good is Qwen3-Next 80B-A3B Instruct?

Qwen3-Next 80B-A3B Instruct by Alibaba (Qwen) ranks 102nd of 354 ranked models on the Noometry Index as of October 2026, with a score of 43.0. Its strongest category is reasoning, where it ranks 81st. API pricing starts at $0.50 per million input tokens and $2 per million output tokens, with a 131K-token context window.

How much does Qwen3-Next 80B-A3B Instruct cost?

Qwen3-Next 80B-A3B Instruct costs $0.50 per million input tokens and $2 per million output tokens on Alibaba (Qwen)'s own API.

What is Qwen3-Next 80B-A3B Instruct's context window?

Qwen3-Next 80B-A3B Instruct accepts up to 131K tokens of input and can write up to 33K tokens in one response.

Is Qwen3-Next 80B-A3B Instruct open source?

Yes. Qwen3-Next 80B-A3B Instruct's weights are downloadable from Hugging Face (Qwen/Qwen3-Next-80B-A3B-Instruct); check the license for commercial terms.

How fast is Qwen3-Next 80B-A3B Instruct?

Qwen3-Next 80B-A3B Instruct generated about 111 output tokens per second in the Kagi LLM Benchmark's timed runs. Speed varies by provider, load and reasoning effort.

What are Qwen3-Next 80B-A3B Instruct's strengths and weaknesses?

Relative to other ranked models, Qwen3-Next 80B-A3B Instruct places best in reasoning, coding, multilingual and lowest in long context, instruction following, writing & preference.

What is Qwen3-Next 80B-A3B Instruct best at?

Its best category is reasoning, where it ranks 81st on Noometry.