File size: 2,003 Bytes
fc196ef
 
 
b93c2dc
fc196ef
 
 
 
 
b93c2dc
 
fc196ef
 
 
 
b93c2dc
 
 
fc196ef
b93c2dc
 
 
 
 
 
 
 
 
b45d5ad
b93c2dc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
---

library_name: transformers
pipeline_tag: text-generation
license: apache-2.0
tags:
  - adaptiq
  - 210-code-labs
  - code
  - agent
  - routed-experts
  - openai-compatible
---


# AdaptIQ-7B

AdaptIQ is an open-weight, self-hosted routed-expert AI system from **210 Code Labs**.
This repository publishes the AdaptIQ model card, manifest, and deployment metadata for
`AdaptIQ-7B`.

## Architecture

AdaptIQ uses expert composition instead of unsafe direct weight blending across incompatible
architectures:

| Capability | Expert |
|---|---|
| Text / reasoning | `mistralai/Mistral-7B-Instruct-v0.3` |
| Code / debugging | `ibm-granite/granite-8b-code-instruct` |
| Image generation | `stabilityai/sd-turbo` |

Composition method: `routed_experts`. `alpha=0.5` is retained only for compatible linear-merge
experiments and is not used by the routed-expert release path.

## Capabilities

- General text chat
- Code generation and debugging
- Agent-oriented reasoning workflows
- Image generation through the routed image expert
- OpenAI-compatible hosted API on Modal
- MCP-ready tool schema scaffold

## Hosted API

Modal endpoint:

```text

https://codelabs--adaptiq-api-api.modal.run/v1

```

Model name:

```text

AdaptIQ-7B

```

Example:

```bash

curl https://codelabs--adaptiq-api-api.modal.run/v1/models \

  -H "Authorization: Bearer $ADAPTIQ_API_KEY"

```

## Intended use

AdaptIQ is intended for software development, automation, research, education, and creative
self-hosted generation workflows.

## Limitations

This release is a routed expert system metadata release. It does not package all upstream expert
weights into a single checkpoint. Runtime deployments download/load the configured open-weight
experts according to the AdaptIQ manifest.

## License

The AdaptIQ release metadata is Apache-2.0. Respect the licenses and model cards of all upstream
experts before redistribution or commercial deployment.