Mind LabLanguage modelWeights available
Macaron-V1-Tall
21out of 100first value
usage data only, range 7 to 35
Summary
Macaron-V1-Tall is Mind Lab's smaller V1 variant, built on Qwen3.6-35B-A3B. Four LoRA specialists and a router cover chat, agent tasks, coding, and Generative UI.
Mind Lab describes the open-weights checkpoint as designed for local deployment and lower latency. Alibaba Cloud Model Studio has offered API access in Singapore since November 16, 2026.
On this page
Popularity compared
How often the model is searched for, used and mentioned, compared with every model of its category.
Popularity index
The 100 most popular models of this category, 0 to 100, higher is more popular
20 of 101 models
As of November 2026 · Source: gradually.ai popularity index
Over time
First value from November 16, 2026. The timeline appears from the second week.
What is gaining popularity right now, across models, vendors, harnesses and AI subscriptions, is collected in All AI trends.
Similar models
More models of the same kind as Macaron-V1-Tall, chosen by model line, developer, size, price and release date.
- Macaron-V1-VentiMind Lab · same developer
- Macaron-V1.1Mind Lab · same developer
- Macaron-V1-Coding-VentiMind Lab · same developer
- Kimi K3Moonshot AI · released around the same time
- GLM-5.3Z.ai · released around the same time
- GLM-5.3-FlashZ.ai · released around the same time
- DeepSeek-V4.1-FlashDeepSeek · released around the same time
- DeepSeek V4 Flash 0731DeepSeek · released around the same time
All specifications
Architecture
- Architecture
- Mixture of LoRA with four specialists for chat, agent tasks, coding, and Generative UI, plus an L0 router.
- Source: Architecture
Input and output
Watermarking and provenance
AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.
- AI watermarking
- Unknown
- no source
- Watermarking technique
- Not known
- no source
Licence and access
- Weights license
- MIT license according to the model card. Any applicable base-model and dependency requirements must also be observed.
- Source: Weights license
- Shutdown
- None announced
- no source
API
Further details
- Parameter-count notes
- The model card describes a Qwen3.6-35B-A3B base plus four LoRA adapters and gives about 50B as the nominal release scale. Mind Lab's V1 announcement lists four 3.7B specialists, while the current Mint API lineup lists four 1B specialists. No total is entered as the model's parameter count.
- Source: Parameter-count notes
- Singapore API context and output limits
- 262,144-token input and context limit, 65,536-token output limit, including thinking mode.
- Source: Singapore API context and output limits
- Regional API availability
- Alibaba Cloud Model Studio lists the API ID in Singapore under its international deployment scope from November 16, 2026.
- Source: Regional API availability
Prices
- Singapore API pricing
- International deployment scope, original rates excluding limited-time promotions: $0.45 input and $2.60 output per 1 million tokens. Implicit cache hits, explicit cache creation with a 5-minute TTL, and explicit cache hits each cost $0.08 per 1 million tokens.Source
Frequently asked questions about Macaron-V1-Tall
How large is the context window of Macaron-V1-Tall?
262,144 tokens. That is how much text Macaron-V1-Tall reads at once, about 390 A4 pages.
Page count as a rule of thumb from Artificial Analysis
Can I run Macaron-V1-Tall locally?
Yes, the weights are open.
Stated by Mind Lab
- Open weights
- yes
- License
- MIT license according to the model card. Any applicable base-model and dependency requirements must also be observed.
- Parameters
- n/a
- active per token
- n/a