Models / Qwen3-Next-80B-A3B-Thinking

Alibaba

Qwen3-Next-80B-A3B-Thinking

Qwen3-Next-80B-A3B-Thinking is strongest at sticking to the facts; it is not on sale through OpenRouter, so it is here for reference.

Results 16 Nov 2026 · 23 results · 4 sources

Price per million tokens
Not available through OpenRouter
Speed
163 tokens a second; on its test prompt the answer starts after 12.7 s, thinking time included
Artificial Analysis, on its usual API
Intelligence Index
11.2
Artificial Analysis
Spring Prompt overall
Not ranked yet
needs our benchmarks and two groups of results
Developer
Alibaba, based in China
Weights not published
Released
11 Sep 2025 (Artificial Analysis)

Where it stands among the models you could choose

Available models grouped by price, intelligence, speed and developer location. Retired models are excluded.

AmongIntelligence IndexSpring Prompt overallPrice, blendedSpeed
Models available today 131st of 186Claude Opus 5.5 – – 18th of 66Nemotron 3.5 Super VL
Models with similar intelligence sets this group – – 6th of 14gpt-oss-20b
Models writing 100–200 tokens a second 20th of 22Claude Sonnet 5.5 – – sets this group
Models from developers based in China 58th of 76MiMo-V2.6-Pro – – 4th of 24DeepSeek-V4.1-Flash

The name under each rank is the leader of that group; green is the top quarter of the group and red the bottom quarter, by rank. Price is shown as a percentile: the share of the group that costs less, so lower is cheaper. "Sets this group" marks the measure the group is defined by. Spring Prompt overall is our score out of 100 across every source (how it works). Price is the developer's list price (or the typical OpenRouter host where we haven't read one) for three input tokens to one output; intelligence and speed are from Artificial Analysis (data sourced from Artificial Analysis); "similar intelligence" means within 4 points on its Intelligence Index. Developer locations are where each company is based, not where a model is served. Groups under 3 models are not ranked.

How good is it, and for what?

Each use case ranks the available models its benchmarks measured. The bar shows its percentile in that field, best to the right. Open a row for the results behind it.

Product listingsTurning a sparse product feed and photos into listings that can go live Not measured
  • CatalogBench: has not measured this model.
Decks from an analysisTurning a finished analysis into a deck you could present as it is Not measured
  • DeckBench: has not measured this model.
User surveysPlanning a user survey and reading its results without being misled Not measured
  • SurveyBench: has not measured this model.
Marketing planningPlanning a year of ad spend without overspending Not measured
  • ROASBench: has not measured this model.
Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls 141st of 160 available
BenchmarkQwen3-Next-80B-A3B-ThinkingRank among availableBest available
Artificial AnalysisArtificial Analysis: Terminal-Bench 2.1 · reported 6.7% 93rd of 108 Claude Fable 5.1 91.4%
Artificial AnalysisArtificial Analysis: τ-bench banking · reported 6.2% 85th of 105 Grok 4.6 50.7%
  • tau2-bench: has not measured this model.
  • Berkeley Function Calling Leaderboard (BFCL) V4: has not measured this model.
  • OpenHands Index: has not measured this model.
  • Microsoft STATE-Bench: has not measured this model.
  • Vending-Bench 2: has not measured this model.
Professional workReal tasks from banking, consulting and law, business documents and freelance projects Not measured
  • APEX-Agents: has not measured this model.
  • GDP.pdf: has not measured this model.
  • Remote Labor Index: has not measured this model.
Reasoning and knowledgeHard questions across science, maths and general knowledge 126th of 188 available
BenchmarkQwen3-Next-80B-A3B-ThinkingRank among availableBest available
Artificial AnalysisArtificial Analysis: Artificial Analysis Intelligence Index · reported 11.2 131st of 186 Claude Opus 5.5 57.6
Artificial AnalysisArtificial Analysis: Humanity's Last Exam · reported 12.6% 120th of 187 Claude Opus 5.5 61.4%
Artificial AnalysisArtificial Analysis: GPQA Diamond · reported 75.9% 107th of 169 GPT-6 Astra 96.3%
WritingWhat people prefer in blind comparisons, and judged writing quality 133rd of 170 available
BenchmarkQwen3-Next-80B-A3B-ThinkingRank among availableBest available
Arena (formerly LMArena)Arena Text: overall · reported 1,369 112th of 154 Gemini 4 Argon 1,525
UGI LeaderboardUGI: writing score · reported 27.2 110th of 125 Gemini 3.8 Flash 78.6
Sticking to the factsSummarising without inventing things, and factual answers 64th of 140 available
BenchmarkQwen3-Next-80B-A3B-ThinkingRank among availableBest available
Vectara Hallucination LeaderboardVectara: hallucination rate · reported 9.3% 34th of 72 GPT-5.4 nano 3.1%
  • Arena (formerly LMArena): has not measured this model.
  • SimpleQA Verified (Epoch AI): has not measured this model.
Speed under pressureGood decisions against a real clock (fast chess) Not measured
  • BulletBench: has not measured this model.

Nearest alternatives

Against Qwen3-Next-80B-A3B-Thinking, Intelligence Index 11.2. "Similar intelligence" means within 4 points or better.

Against the alternatives

The models you would most likely weigh it against: the leaders of the groups above.

Scroll sideways to see every alternative.

Qwen3-Next-80B-A3B-Thinkinggpt-oss-20b
fastest with similar intelligence
Claude Opus 5.5
most intelligent available today
MiMo-V2.6-Pro
most intelligent from China
Price per million tokens – $0.06$8.00$0.54
Tokens a second 163 2779643
Intelligence Index 11.2 10.057.646.3
Spring Prompt overall – –82–
Rank among available models, by use case
Product listings – –13th–
Decks from an analysis – –7th–
User surveys – –4th–
Marketing planning – –9th–
Agents and tool use 141st of 160 129th3rd23rd
Professional work – –3rd–
Reasoning and knowledge 126th of 188 133rd1st13th
Writing 133rd of 170 156th5th23rd
Sticking to the facts 64th of 140 –3rd10th
Speed under pressure – –––

Shaded figures are better than Qwen3-Next-80B-A3B-Thinking's. A dash means no result.

Where it has been measured

Compared with the models available today (it is not on sale itself), in the bottom quarter for agents and tool use, and writing.

Every result

23 published results from 4 sources, each in the source's own units, with the configuration that produced it.

Arena (formerly LMArena) · reported by Arena (formerly LMArena) · 6 results
MeasureValueRankConfigurationDated
Arena Text: business, management and financeArena rating, higher is better 1,367
1,355–1,378
197 of 406 Qwen3-Next-80B-A3B-Thinking 2 Oct 2026
Arena Text: creative writingArena rating, higher is better 1,326
1,312–1,341
204 of 411 Qwen3-Next-80B-A3B-Thinking 2 Oct 2026
Arena Text: expert promptsArena rating, higher is better 1,393
1,370–1,416
187 of 363 Qwen3-Next-80B-A3B-Thinking 2 Oct 2026
Arena Text: instruction followingArena rating, higher is better 1,359
1,350–1,369
195 of 413 Qwen3-Next-80B-A3B-Thinking 2 Oct 2026
Arena Text: overallArena rating, higher is better 1,369
1,363–1,375
195 of 413 Qwen3-Next-80B-A3B-Thinking 2 Oct 2026
Arena Text: writing, literature and languageArena rating, higher is better 1,341
1,331–1,352
202 of 412 Qwen3-Next-80B-A3B-Thinking 2 Oct 2026

Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Artificial Analysis · reported by Artificial Analysis · 12 results
MeasureValueRankConfigurationDated
Artificial Analysis: Artificial Analysis Coding Indexindex score, higher is better 17.4 160 of 187 Qwen3-Next-80B-A3B-Thinking 20 Oct 2026
Artificial Analysis: Artificial Analysis Intelligence Indexindex score, higher is better 11.2 303 of 413 Qwen3-Next-80B-A3B-Thinking 20 Oct 2026
Artificial Analysis: GPQA Diamond% of questions, higher is better 75.9% 212 of 360 Qwen3-Next-80B-A3B-Thinking 20 Oct 2026
Artificial Analysis: Humanity's Last Exam% of questions, higher is better 12.6% 253 of 419 Qwen3-Next-80B-A3B-Thinking 20 Oct 2026
Artificial Analysis: IFBench% of instructions, higher is better 60.7% 104 of 288 Qwen3-Next-80B-A3B-Thinking 20 Oct 2026
Artificial Analysis: Long-context reasoning (AA-LCR)% of questions, higher is better 63.7% 240 of 407 Qwen3-Next-80B-A3B-Thinking 20 Oct 2026
Artificial Analysis: Output speedtokens per second, higher is better 163 30 of 137 Qwen3-Next-80B-A3B-Thinking 20 Oct 2026
Artificial Analysis: Terminal-Bench 2.1% of tasks, higher is better 6.7% 163 of 188 Qwen3-Next-80B-A3B-Thinking 20 Oct 2026
Artificial Analysis: Terminal-Bench Hard% of tasks, higher is better 9.8% 202 of 282 Qwen3-Next-80B-A3B-Thinking 20 Oct 2026
Artificial Analysis: Time to first answer tokenseconds, lower is better 12.7 65 of 137 Qwen3-Next-80B-A3B-Thinking 20 Oct 2026
Artificial Analysis: τ-bench banking% of tasks, higher is better 6.2% 146 of 175 Qwen3-Next-80B-A3B-Thinking 20 Oct 2026
Artificial Analysis: τ²-bench telecom% of tasks, higher is better 41.5% 178 of 286 Qwen3-Next-80B-A3B-Thinking 20 Oct 2026

Not shown: Not your codebase or tools.

UGI Leaderboard · reported by UGI Leaderboard · 3 results
MeasureValueRankConfigurationDated
UGI: requested-length error% off the requested word count, lower is better 67.0% 338 of 379 Qwen3-Next-80B-A3B-Thinking 29 Nov 2025
UGI: style adherencescore from 0 to 1, higher is better 0.29 363 of 379 Qwen3-Next-80B-A3B-Thinking 29 Nov 2025
UGI: writing scorescore out of 100, higher is better 27.2 300 of 379 Qwen3-Next-80B-A3B-Thinking 29 Nov 2025

Not shown: Not other format limits such as character counts or bullet counts.

Vectara Hallucination Leaderboard · reported by Vectara Hallucination Leaderboard · 2 results
MeasureValueRankConfigurationDated
Vectara: answer rate% of documents, higher is better 94.4% 95 of 108 Qwen3-Next-80B-A3B-Thinking 22 Sep 2026
Vectara: hallucination rate% of summaries, lower is better 9.3% 47 of 108 Qwen3-Next-80B-A3B-Thinking 22 Sep 2026

Not shown: Not a quality score: a low rate usually means content filters were triggered, and hallucination rates are measured on answered documents only.

Test your own prompt.

Benchmarks aren't your task. Run your prompt on the models you could use, with every answer checked and costed.