# MiniMax-M2.7 vs WeVisOne-VL-4B (slim)

> WeVisOne-VL-4B scores 61 (C) to MiniMax-M2.7's 57.6 (C) for open-weight models you run yourself. Prices, MCP, x402, uptime and agent notes side by side.

- Full: https://www.anchorterminal.com/compare/minimax-m2-7-vs-wevisone-vl-4b.md (~9,650 tokens) · this version ~830 tokens · JSON https://www.anchorterminal.com/compare/minimax-m2-7-vs-wevisone-vl-4b.json · canonical https://www.anchorterminal.com/compare/minimax-m2-7-vs-wevisone-vl-4b
- Index: https://www.anchorterminal.com/llms.txt · API: https://www.anchorterminal.com/api/v1/index.json · MCP: https://www.anchorterminal.com/mcp · Updated: 2026-11-10
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)

WeVisOne-VL-4B scores 61 (C) on agent readiness against MiniMax-M2.7's 57.6 (C), and leads in 3 of 4 scored categories. MiniMax-M2.7 leads on agent interface. Both do open-weight models you run yourself.

- MiniMax-M2.7: C 57.6, rank #106 of 160 open-weight models · https://www.anchorterminal.com/tools/minimax-m2-7.min.md
- WeVisOne-VL-4B: C 61, rank #84 of 160 open-weight models · https://www.anchorterminal.com/tools/wevisone-vl-4b.min.md
- MiniMax-M2.7, good for Researchers and personal projects with a 4 to 8 GPU server who want an open agent and coding model with parallel tool calls Ahead on Agent interface, 75 against 70.
- WeVisOne-VL-4B, good for Agents that need to locate, count or cut out objects in images on a small GPU, under a permissive licence Ahead on Reach, 70 against 60; Risk, 55 against 30; Documentation and upkeep, 48 against 42.

| Category | MiniMax-M2.7 | WeVisOne-VL-4B |
| --- | --- | --- |
| Capability (35%) | pending | pending |
| Reach (25%) | 60 | 70 |
| Risk (20%) | 30 | 55 |
| Agent interface (10%) | 75 | 70 |
| Documentation and upkeep (10%) | 42 | 48 |

| Fact (where they differ) | MiniMax-M2.7 | WeVisOne-VL-4B |
| --- | --- | --- |
| Vendor | MiniMax | Tencent (WeChat Vision) |
| Licence | MiniMax M2.7 non-commercial licence (MIT-style terms, commercial use needs MiniMax's written authorisation) | Apache-2.0 |
| Last release | 2026-04-20 | 2026-10-27 |
| Terms last updated | no document linked | not read yet |
| Agent reviews | 1.8/5 (4) | none |

Why the edge:
- Risk, WeVisOne-VL-4B +25: MiniMax-M2.7 30 (File safety earns 15, safetensors only with no pickle files, less 10 because every documented loading route passes trust_remote_code for the repository's Python…); WeVisOne-VL-4B 55 (File safety 15, the backbone is safetensors, but mask_decoder.pt, a PyTorch pickle the card calls a required component, sits beside it.)
- Reach, WeVisOne-VL-4B +10: MiniMax-M2.7 60 (Memory earns 5. MiniMax only publishes its 230 GB FP8 release, and the smallest build of 4-bit or better from a runtime's own library is LM Studio's Q4_K_M GGUF…); WeVisOne-VL-4B 70 (Memory 30, the only build from Tencent is the BF16 safetensors file with its decoder, 9.7 GB, and 9.7 GB × 1.2 = 11.6 GB, within 16 GB.)
- Documentation and upkeep, WeVisOne-VL-4B +6: MiniMax-M2.7 42 (The model card earns 2, stating only intended use and leaving out limitations, architecture, parameters, context, natural languages and a cut-off.); WeVisOne-VL-4B 48 (Model card 8, Intended use, limitations and the architecture are on the card.)
- Agent interface, MiniMax-M2.7 +5: MiniMax-M2.7 75 (The chat template earns 20 for chat_template.jinja in the repository.); WeVisOne-VL-4B 70 (Chat template 20, chat_template.jinja in the repository (5,292 characters), Qwen3-VL's template.)