Compare / Alibaba vs Anthropic

AI labs compared · November 2026

Alibaba vs Anthropic: which AI lab is ahead?

Each lab's best model on sale goes head to head on every kind of work we measure, from our own benchmarks and licensed sources. 38 models in all; a lab is only as good as the model you would actually pick from it.

22 models on sale with our resultsAlibaba Best: Qwen4-Max · newest: Qwen4-Max
16 models on sale with our resultsAnthropic Best: Claude Fable 5.2 · newest: Claude Fable 5.2
Alibaba wins 1Anthropic wins 9

Kinds of work led by each lab's best model there, of the 10 both are measured on.

Results as of 19 November 2026 (catalogue release 2026-11-19-3608133e8331). Models on sale means listed on OpenRouter today.

Alibaba or Anthropic? The short answer

Anthropic is ahead: its best models beat Alibaba's in 9 of the 10 kinds of work both are measured on; Alibaba leads in 1.

  • Pick Alibaba for decks from an analysis.
  • Pick Anthropic for product listings, user surveys, marketing planning, agents and tool use, professional work, reasoning and knowledge, writing, sticking to the facts and speed under pressure.
  • On our overall leaderboard, Claude Fable 5.2 is #1 (88 of 100) and Qwen4-Max is #19 (53).
  • Prices start at $0.11 per million tokens for Alibaba (Qwen3.5-9B) and $0.20 for Anthropic (Claude Haiku 5.5).

Head to head

53Qwen4-Max
Best Spring Prompt overall score
88Claude Fable 5.2
45.4Qwen3.8-Max (0902)
Best Intelligence Index
57.6Claude Opus 5.5
$2.25Qwen4-Max
Flagship price, per million tokens
$16.00Claude Fable 5.2
$0.11Qwen3.5-9B
Cheapest model, per million tokens
$0.20Claude Haiku 5.5
22
Models on sale with our results
16

Each line names the model behind the number. Prices are blended, three input tokens to one output, at the developer's list price where we have read it; see every model's price. Intelligence Index from Artificial Analysis; overall score from our leaderboard.

Which lab is better for what

For each kind of work, each lab's best model there and its rank among every model on sale. Further right is better; the better rank is in green.

Product listingsTurning a sparse product feed and photos into listings that can go live 22 of 35Qwen4-Max 8 of 35Claude Fable 5.2
Decks from an analysisTurning a finished analysis into a deck you could present as it is 1 of 35Qwen4-Max 4 of 35Claude Fable 5.2
User surveysPlanning a user survey and reading its results without being misled 6 of 26Qwen4-Max 1 of 26Claude Fable 5.2
Marketing planningPlanning a year of ad spend without overspending 18 of 35Qwen4-Max 1 of 35Claude Fable 5.2
Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls 10 of 159Qwen3.8-Max (0902) 2 of 159Claude Sonnet 5.5
Professional workReal tasks from banking, consulting and law, business documents and freelance projects 18 of 51Qwen3.8-Max (0902) 1 of 51Claude Sonnet 5.5
Reasoning and knowledgeHard questions across science, maths and general knowledge 21 of 187Qwen3.8-Max (0902) 1 of 187Claude Opus 5.5
WritingWhat people prefer in blind comparisons, and judged writing quality 17 of 169Qwen3.8-Max (0902) 2 of 169Claude Fable 5
Sticking to the factsSummarising without inventing things, and factual answers 14 of 139Qwen3-14B 3 of 139Claude Opus 5.5
Speed under pressureGood decisions against a real clock (fast chess) 29 of 38Qwen4-Max 10 of 38Claude Fable 5.2

Every model, by intelligence and price

Each dot is a model on sale with an Intelligence Index from Artificial Analysis. Up is smarter; left is cheaper. Hover or tap a dot for its name.

AlibabaAnthropic

020406080 $0.1$1$10$100 Price per million tokens, blended (log scale) Intelligence Index Qwen3.8-Max (0902)Claude Opus 5.5Claude Fable 5.1Claude Sonnet 5.5Claude Opus 5Claude Fable 5

On our overall leaderboard

The 8 of their models we rank overall: one score out of 100 across everything we publish (how the score works). Only models with enough of our own results are ranked.

  1. #1Claude Fable 5.288
  2. #4Claude Opus 5.580
  3. #12Claude Fable 5.164
  4. #14Claude Sonnet 5.560
  5. #19Qwen4-Max53
  6. #23Qwen3.8-Max (0902)44
  7. #27Claude Haiku 5.538
  8. #39Claude Haiku 4.57

Benchmark by benchmark

The headline result of every benchmark both labs have models on: each lab's best model and its rank among the models on sale. Green marks the better rank.

MeasureAlibabaAnthropic
Agents and tool use
τ²-bench · Task success (pass^1) · tau2 v1.0.1, retail% of tasks, higher is better 84.4%Qwen3.5-397B-A17B · 1 of 679.6%Claude Opus 4.5 · 3 of 6
τ²-bench · Task success (pass^1) · tau2 v1.0.1, airline% of tasks, higher is better 81.5%Qwen3.5-397B-A17B · 5 of 684.0%Claude Opus 4.5 · 1 of 6
τ²-bench · Task success (pass^1) · tau2 v1.0.1, telecom% of tasks, higher is better 97.8%Qwen3.5-397B-A17B · 1 of 692.3%Claude Opus 4.5 · 2 of 6
Berkeley Function Calling Leaderboard (BFCL) V4 · Overall accuracy · BFCL% correct, higher is better 52.2%Qwen3-235B-A22B-Instruct-2507 · 12 of 3677.5%Claude Opus 4.5 · 1 of 36
OpenHands Index · Average% resolved, higher is better 52.9%Qwen3.6-Plus · 16 of 3181.0%Claude Fable 5 · 1 of 31
Artificial Analysis · Terminal-Bench 4.0% of tasks, higher is better 38.9%Qwen3.8-Max (0902) · 15 of 10763.6%Claude Sonnet 5.5 · 2 of 107
Artificial Analysis · Terminal-Bench 2.1% of tasks, higher is better 88.8%Qwen3.8-Max (0902) · 5 of 10791.4%Claude Fable 5.1 · 1 of 107
Artificial Analysis · τ-bench banking% of tasks, higher is better 48.0%Qwen3.8-27B · 5 of 10447.2%Claude Fable 5.1 · 7 of 104
Vending-Bench 2 · Money after a yearUS dollars, higher is better $5,115Qwen3.6-Plus · 24 of 53$11,182Claude Opus 5 · 3 of 53
Decks from an analysis
DeckBench · Deck ratingrating, higher is better 1,563Qwen4-Max · 1 of 351,403Claude Fable 5.2 · 4 of 35
Marketing planning
ROASBench · Overall scorescore out of 100, higher is better 45.8Qwen4-Max · 18 of 3561.3Claude Fable 5.2 · 1 of 35
Product listings
CatalogBench · Reliably publish-ready% of products, higher is better 39.3%Qwen4-Max · 22 of 3571.4%Claude Fable 5.2 · 6 of 35
CatalogBench · Reliably publish-ready · sales brief% of products, higher is better 8.9%Qwen4-Max · 24 of 3544.6%Claude Fable 5.2 · 9 of 35
Professional work
APEX-Agents · Tasks passed% of tasks, higher is better 24.9%Qwen3.5-397B-A17B · 36 of 3775.5%Claude Sonnet 5.5 · 1 of 37
GDP.pdf · Professional document tasks% of rubric, higher is better 23.2%Qwen3.8-Max (0902) · 13 of 3130.6%Claude Opus 5.5 · 2 of 31
Reasoning and knowledge
Artificial Analysis · Artificial Analysis Intelligence Indexindex score, higher is better 45.4Qwen3.8-Max (0902) · 15 of 18557.6Claude Opus 5.5 · 1 of 185
Artificial Analysis · Humanity's Last Exam% of questions, higher is better 43.1%Qwen3.8-Max (0902) · 30 of 18661.4%Claude Opus 5.5 · 1 of 186
Artificial Analysis · GPQA Diamond% of questions, higher is better 92.8%Qwen3.8-Max (0902) · 15 of 16893.7%Claude Fable 5.1 · 8 of 168
Speed under pressure
BulletBench · Ladder Elo · Bullet 60sladder Elo, higher is better 147Qwen4-Max · 28 of 36491Claude Fable 5.2 · 13 of 36
Sticking to the facts
Vectara Hallucination Leaderboard · Hallucination rate · Vectara% of summaries, lower is better 5.4%Qwen3-14B · 10 of 719.8%Claude Haiku 4.5 · 40 of 71
Arena (formerly LMArena) · Overall · Text factualityArena rating, higher is better 1,472Qwen3.8-Max (0902) · 19 of 1091,507Claude Opus 5.5 · 2 of 109
SimpleQA Verified (Epoch AI) · Correct answers% of questions, higher is better 55.8%Qwen3.7-Max · 18 of 6872.2%Claude Opus 5.5 · 4 of 68
User surveys
SurveyBench · Analysis ratingrating, higher is better 1,381Qwen4-Max · 6 of 261,670Claude Fable 5.2 · 1 of 26
Writing
Arena (formerly LMArena) · Overall · TextArena rating, higher is better 1,482Qwen3.8-Max (0902) · 18 of 1531,505Claude Opus 4.6 · 2 of 153
UGI Leaderboard · Writing score · UGIscore out of 100, higher is better 58.2Qwen3.6-Plus · 36 of 12474.2Claude Fable 5 · 3 of 124

Their models, one against one

Quick answers

Is Alibaba or Anthropic better?

Anthropic is ahead: its best models beat Alibaba's in 9 of the 10 kinds of work both are measured on; Alibaba leads in 1.

What is Alibaba's best model?

Qwen4-Max: #19 on the Spring Prompt overall leaderboard.

What is Anthropic's best model?

Claude Fable 5.2: #1 on the Spring Prompt overall leaderboard.

Which is better for agents and tool use, Alibaba or Anthropic?

Anthropic: its best model there, Claude Sonnet 5.5, ranks 2 of 159 models on sale, against 10 for Alibaba's Qwen3.8-Max (0902).

Which is better for reasoning and knowledge, Alibaba or Anthropic?

Anthropic: its best model there, Claude Opus 5.5, ranks 1 of 187 models on sale, against 21 for Alibaba's Qwen3.8-Max (0902).

Which is better for writing, Alibaba or Anthropic?

Anthropic: its best model there, Claude Fable 5, ranks 2 of 169 models on sale, against 17 for Alibaba's Qwen3.8-Max (0902).

Which is cheaper, Alibaba or Anthropic?

Alibaba's cheapest model with results, Qwen3.5-9B, costs $0.11 per million tokens (three input to one output), against $0.20 for Claude Haiku 5.5 from Anthropic.

Other labs compared

Which model is best for your product?

Benchmarks aren't your task. Run models from Alibaba, Anthropic and the rest on your own prompt, and see the cheapest one that passes.