ApertureAug 2026
FieldCompareLabsGuideMethod
  • Field
  • Compare
  • Labs
  • Guide
Q

Alibaba (Qwen)

Hangzhou

Qwen3.8 Max is the best evidence-backed open-weight flagship. Strong multilingual and reasoning; slower on tool loops.

qwen.ai
01
Q

Qwen3.8 Max

Alibaba (Qwen) · Aug 2026 · Open weight

58.1
Index
Intel58.1
Code71.8
Agents58.4

Best open-weight flagship on evidence-backed boards. Slow to sample. Strong multilingual and agentic for the license.

Price /1M
$2 / $6
Context
1M
Speed
21 t/s
Status
Current
Open-weight frontierSelf-host at scale
Details
02
Q

Qwen3.8 27B

Alibaba (Qwen) · Aug 2026 · Open weight

52
Index
Intel52
Code68.1
Agents50.9

Dense 27B that hangs with models 50× its size. The local / workstation Qwen, not the datacenter Max.

Price /1M
$2.55 / $2.55
Context
262K
Speed
47 t/s
Status
Current
Local 24–80GB boxesFine-tunes
Details
03
Q

Qwen3.8 Flash

Alibaba (Qwen) · Aug 2026 · Open weight

50
Index
Intel50
Code64
Agents44

Qwen’s cheap Fast lane. Use it as the open-weight Luna, not as a Max replacement.

Price /1M
$0.16 / $0.47
Context
1M
Speed
90 t/s
Status
Current
Open-weight volumeBatch
Details

Aperture is a field guide, not a vendor. Scores compiled 28 Aug 2026.

MethodologyHow to pick