Tier
Major
Headquarters
Hangzhou, China
Founded
1999
Flagship
Qwen
Models tracked
81
Artificial Analysis
Top intelligence
56.6
Artificial Analysis Index
Cheapest model
$0.00
blended /1M
Fastest output
339
tokens / sec
Flagship
Alibaba’s top-scoring model on the Artificial Analysis Intelligence Index.
Most intelligent model
Qwen3.7 Max
Released 2026-05-19
Intelligence
56.6
Coding
50.1
Math
—
In $/1M
$2.50
Out $/1M
$7.50
Speed t/s
105
Benchmark breakdown
Models & benchmarks
Every Alibaba model on file — intelligence, coding, price, and speed.
| Model | Released | Intel | Coding | Math | In $/1M | Out $/1M | Speed t/s | Latency s |
|---|---|---|---|---|---|---|---|---|
| Qwen3.7 Max | 2026-05-19 | 56.6 | 50.1 | — | $2.50 | $7.50 | 105 | 1.72 |
| Qwen3.7 Plus | 2026-06-01 | 53.3 | 46.5 | — | $0.40 | $1.16 | 53 | 1.34 |
| Qwen3.6 Max Preview | 2026-04-20 | 51.8 | 44.9 | — | $1.30 | $7.80 | 41 | 1.99 |
| Qwen3.6 Plus | 2026-04-02 | 50.0 | 42.9 | — | $0.50 | $3.00 | 53 | 1.75 |
| Qwen3.6 27B (Reasoning) | 2026-04-22 | 45.8 | 36.5 | — | $0.60 | $3.60 | 66 | 1.45 |
| Qwen3.5 397B A17B (Reasoning) | 2026-02-16 | 45.0 | 41.3 | — | $0.60 | $3.60 | 52 | 1.88 |
| Qwen3.6 35B A3B (Reasoning) | 2026-04-16 | 43.5 | 35.2 | — | $0.25 | $1.49 | 171 | 1.35 |
| Qwen3.5 27B (Reasoning) | 2026-02-24 | 42.1 | 34.9 | — | $0.30 | $2.40 | 84 | 1.37 |
| Qwen3.5 122B A10B (Reasoning) | 2026-02-24 | 41.6 | 34.7 | — | $0.40 | $3.20 | 137 | 1.17 |
| Qwen3.5 397B A17B (Non-reasoning) | 2026-02-16 | 40.1 | 37.4 | — | $0.60 | $3.60 | 53 | 1.86 |
| Qwen3 Max Thinking | 2026-01-26 | 39.8 | 30.5 | — | $1.20 | $6.00 | 0 | 0.00 |
| Qwen3.5 Omni Plus | 2026-03-30 | 38.6 | 27.6 | — | $0.40 | $4.80 | 50 | 1.30 |
| Qwen3.5 27B (Non-reasoning) | 2026-02-24 | 37.2 | 33.4 | — | $0.30 | $2.60 | 91 | 1.38 |
| Qwen3.5 35B A3B (Reasoning) | 2026-02-24 | 37.1 | 30.3 | — | $0.25 | $2.00 | 145 | 1.25 |
Trajectory
Alibaba’s model releases over time, by intelligence.
House strengths
Where Alibaba’s lineup over-indexes, averaged across its models.
Capability profile
Best value
Qwen3.5 0.8B (Reasoning)
Most intelligence per dollar — 10.5 index at $0.02/1M blended.
525 index points per $1
Media models
Image, video, and speech models, ranked by arena Elo.
| Model | Type | Released | Elo | Rank |
|---|---|---|---|---|
| Wan 2.5 Preview | image to video | 2025-09 | 1237 | #35 |
| Fun-Realtime-TTS | text to speech | — | 1231 | #1 |
| Wan 2.6 | image to video | 2025-12 | 1209 | #44 |
| Wan 2.7 Pro | image editing | 2026-04 | 1206 | #16 |
| Wan 2.6 Image | image editing | 2025-12 | 1197 | #17 |
| Wan 2.6 | text to video | 2025-12 | 1196 | #29 |
| Wan 2.7 | image editing | 2026-04 | 1189 | #22 |
| Qwen Image 2.0 Pro (2026-04-22) | text to image | 2026-04 | 1174 | #22 |
| Qwen Image 2.0 Pro (2026-04-22) | image editing | 2026-04 | 1171 | #26 |
| Qwen Image Edit Max 2601 | image editing | 2026-01 | 1162 | #32 |
| Qwen Image Max 2512 | text to image | 2025-12 | 1161 | #28 |
| Wan 2.5 Preview | text to video | 2025-09 | 1159 | #42 |
Overview
The cloud and AI arm of the Alibaba Group, whose Qwen family is among the most prolific open-weight model lines in the world, spanning dense and mixture-of-experts models across text, vision, and code.
Visit Alibaba