You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/cli/configuration/settings.mdx
+4-2Lines changed: 4 additions & 2 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -27,7 +27,7 @@ If the file doesn't exist, it's created with defaults the first time you run **d
27
27
28
28
| Setting | Options | Default | Description |
29
29
| ------- | ------- | ------- | ----------- |
30
-
|`model`|`opus`, `opus-4-6`, `opus-4-6-fast`, `sonnet`, `gpt-5.1`, `gpt-5.1-codex`, `gpt-5.1-codex-max`, `gpt-5.2`, `gpt-5.2-codex`, `haiku`, `gemini-3-pro`, `droid-core`, `kimi-k2.5`, `custom-model`|`opus`| The default AI model used by droid |
30
+
|`model`|`opus`, `opus-4-6`, `opus-4-6-fast`, `sonnet`, `gpt-5.1`, `gpt-5.1-codex`, `gpt-5.1-codex-max`, `gpt-5.2`, `gpt-5.2-codex`, `gpt-5.3-codex`, `haiku`, `gemini-3-pro`, `droid-core`, `kimi-k2.5`, `minimax-m2.5`, `custom-model`|`opus`| The default AI model used by droid |
31
31
|`reasoningEffort`|`off`, `none`, `low`, `medium`, `high` (availability depends on the model) | Model-dependent default | Controls how much structured thinking the model performs. |
32
32
|`autonomyLevel`|`normal`, `spec`, `auto-low`, `auto-medium`, `auto-high`|`normal`| Sets the default autonomy mode when starting droid. |
33
33
|`cloudSessionSync`|`true`, `false`|`true`| Mirror CLI sessions to Factory web. |
@@ -62,11 +62,13 @@ Choose the default AI model that powers your droid:
62
62
-**`gpt-5.1-codex`** - Advanced coding-focused model
63
63
-**`gpt-5.1-codex-max`** - GPT-5.1-Codex-Max, supports Extra High reasoning
64
64
-**`gpt-5.2`** - OpenAI GPT-5.2
65
-
-**`gpt-5.2-codex`** - GPT-5.2-Codex, latest OpenAI coding model with Extra High reasoning
65
+
-**`gpt-5.2-codex`** - GPT-5.2-Codex, OpenAI coding model with Extra High reasoning
66
+
-**`gpt-5.3-codex`** - GPT-5.3-Codex, latest OpenAI coding model with Extra High reasoning and verbosity support
66
67
-**`haiku`** - Claude Haiku 4.5, fast and cost-effective
67
68
-**`gemini-3-pro`** - Gemini 3 Pro
68
69
-**`droid-core`** - GLM-4.7 open-source model
69
70
-**`kimi-k2.5`** - Kimi K2.5 open-source model with image support
71
+
-**`minimax-m2.5`** - MiniMax M2.5 open-source model with reasoning support (0.12× multiplier)
70
72
-**`custom-model`** - Your own configured model via BYOK
71
73
72
74
[You can also add custom models and BYOK.](/cli/configuration/byok)
Model quality evolves quickly, and we tune the CLI defaults as the ecosystem shifts. Use this guide as a snapshot of how the major options compare today, and expect to revisit it as we publish updates. This guide was last updated on Thursday, February 12th 2026.
7
+
Model quality evolves quickly, and we tune the CLI defaults as the ecosystem shifts. Use this guide as a snapshot of how the major options compare today, and expect to revisit it as we publish updates. This guide was last updated on Friday, February 14th 2026.
8
8
9
9
---
10
10
@@ -17,15 +17,17 @@ Model quality evolves quickly, and we tune the CLI defaults as the ecosystem shi
17
17
| 3 |**Claude Opus 4.5**| Proven quality-and-safety balance; strong default for TUI and exec. |
18
18
| 4 |**GPT-5.1-Codex-Max**| Fast coding loops with support up to **Extra High** reasoning; great for heavy implementation and debugging. |
19
19
| 5 |**Claude Sonnet 4.5**| Strong daily driver with balanced cost/quality; great general-purpose choice when you don’t need Opus-level depth. |
20
-
| 6 |**GPT-5.2-Codex**| Latest OpenAI coding model with **Extra High** reasoning; strong for implementation-heavy tasks. |
21
-
| 7 |**GPT-5.1-Codex**| Quick iteration with solid code quality at lower cost; bump reasoning when you need more depth. |
22
-
| 8 |**GPT-5.1**| Good generalist, especially when you want OpenAI ergonomics with flexible reasoning effort. |
23
-
| 9 |**GPT-5.2**| Advanced OpenAI model with verbosity support and reasoning up to **Extra High**. |
24
-
| 10 |**Claude Haiku 4.5**| Fast, cost-efficient for routine tasks and high-volume automation. |
25
-
| 11 |**Gemini 3 Pro**| Strong at mixed reasoning with Low/High settings; helpful for researchy flows with structured outputs. |
26
-
| 12 |**Gemini 3 Flash**| Fast, cheap (0.2× multiplier) with full reasoning support; great for high-volume tasks where speed matters. |
27
-
| 13 |**Droid Core (GLM-4.7)**| Open-source, 0.25× multiplier, great for bulk automation or air-gapped environments; note: no image support. |
28
-
| 14 |**Droid Core (Kimi K2.5)**| Open-source, 0.25× multiplier with image support; good for cost-sensitive work. |
20
+
| 6 |**GPT-5.3-Codex**| Newest OpenAI coding model with **Extra High** reasoning and verbosity support; strong for implementation-heavy tasks. |
21
+
| 7 |**GPT-5.2-Codex**| Proven OpenAI coding model with **Extra High** reasoning; solid for implementation-heavy tasks. |
22
+
| 8 |**GPT-5.1-Codex**| Quick iteration with solid code quality at lower cost; bump reasoning when you need more depth. |
23
+
| 9 |**GPT-5.1**| Good generalist, especially when you want OpenAI ergonomics with flexible reasoning effort. |
24
+
| 10 |**GPT-5.2**| Advanced OpenAI model with verbosity support and reasoning up to **Extra High**. |
25
+
| 11 |**Claude Haiku 4.5**| Fast, cost-efficient for routine tasks and high-volume automation. |
26
+
| 12 |**Gemini 3 Pro**| Strong at mixed reasoning with Low/High settings; helpful for researchy flows with structured outputs. |
27
+
| 13 |**Gemini 3 Flash**| Fast, cheap (0.2× multiplier) with full reasoning support; great for high-volume tasks where speed matters. |
28
+
| 14 |**Droid Core (MiniMax M2.5)**| Open-source, 0.12× multiplier with reasoning support (Low/Medium/High); cheapest model available. No image support. |
29
+
| 15 |**Droid Core (GLM-4.7)**| Open-source, 0.25× multiplier, great for bulk automation or air-gapped environments; note: no image support. |
30
+
| 16 |**Droid Core (Kimi K2.5)**| Open-source, 0.25× multiplier with image support; good for cost-sensitive work. |
29
31
30
32
<Note>
31
33
We ship model updates regularly. When a new release overtakes the list above,
@@ -39,10 +41,10 @@ Model quality evolves quickly, and we tune the CLI defaults as the ecosystem shi
|**Deep planning, architecture reviews, ambiguous product specs**| Start with **Opus 4.6** for best depth and safety, or **Opus 4.6 Fast** for faster turnaround. Use **Sonnet 4.5** when you want balanced cost/quality, or **Codex/Codex-Max** for faster iteration with reasoning. |
42
-
|**Full-feature development, large refactors**|**Opus 4.6** or **Opus 4.5** for depth and safety. **GPT-5.2-Codex** or **GPT-5.1-Codex-Max** when you need speed plus **Extra High** reasoning; **Sonnet 4.5** for balanced loops. |
43
-
|**Repeatable edits, summarization, boilerplate generation**|**Haiku 4.5** or **Droid Core** for speed and cost. **GPT-5.1 / GPT-5.1-Codex** when you need higher quality or structured outputs. |
44
+
|**Full-feature development, large refactors**|**Opus 4.6** or **Opus 4.5** for depth and safety. **GPT-5.3-Codex**, **GPT-5.2-Codex**, or **GPT-5.1-Codex-Max** when you need speed plus **Extra High** reasoning; **Sonnet 4.5** for balanced loops. |
45
+
|**Repeatable edits, summarization, boilerplate generation**|**Haiku 4.5** or **Droid Core**(including **MiniMax M2.5** at 0.12×) for speed and cost. **GPT-5.1 / GPT-5.1-Codex** when you need higher quality or structured outputs. |
44
46
|**CI/CD or automation loops**| Favor **Haiku 4.5** or **Droid Core** for predictable, low-cost throughput. Use **Codex** or **Codex-Max** when automation needs stronger reasoning. |
45
-
|**High-volume automation, frequent quick turns**|**Haiku 4.5** for speedy feedback. **Droid Core** when cost is critical or you need air-gapped deployment. |
47
+
|**High-volume automation, frequent quick turns**|**Haiku 4.5** for speedy feedback. **Droid Core**(especially **MiniMax M2.5** at 0.12× with reasoning) when cost is critical or you need air-gapped deployment. |
46
48
47
49
<Tip>
48
50
**Claude Opus 4.6** is the top-tier option for extremely complex architecture decisions or critical work where you need maximum reasoning capability. **Opus 4.6 Fast** is tuned for faster responses at a higher cost. Most tasks don't require Opus-level power—start with Sonnet 4.5 and escalate only if needed.
@@ -70,12 +72,14 @@ Tip: you can swap models mid-session with `/model` or by toggling in the setting
70
72
-**GPT-5.1-Codex-Max**: Low / Medium / High / **Extra High** (default: Medium)
71
73
-**GPT-5.2**: Off / Low / Medium / High / **Extra High** (default: Low)
72
74
-**GPT-5.2-Codex**: None / Low / Medium / High / **Extra High** (default: Medium)
75
+
-**GPT-5.3-Codex**: None / Low / Medium / High / **Extra High** (default: Medium)
73
76
-**Gemini 3 Pro**: None / Low / Medium / High (default: High)
74
77
-**Gemini 3 Flash**: Minimal / Low / Medium / High (default: High)
75
78
-**Droid Core (GLM-4.7)**: None only (default: None; no image support)
76
79
-**Droid Core (Kimi K2.5)**: None only (default: None)
80
+
-**Droid Core (MiniMax M2.5)**: Low / Medium / High (default: High)
77
81
78
-
Reasoning effort increases latency and cost—start low for simple work and escalate as needed. **Max** is available on Claude Opus 4.6. **Extra High** is available on GPT-5.1-Codex-Max, GPT-5.2, and GPT-5.2-Codex.
82
+
Reasoning effort increases latency and cost—start low for simple work and escalate as needed. **Max** is available on Claude Opus 4.6. **Extra High** is available on GPT-5.1-Codex-Max, GPT-5.2, GPT-5.2-Codex, and GPT-5.3-Codex.
79
83
80
84
<Tip>
81
85
Change reasoning effort from `/model` → **Reasoning effort**, or via the
@@ -90,14 +94,14 @@ Factory ships with managed Anthropic and OpenAI access. If you prefer to run aga
90
94
91
95
### Open-source models
92
96
93
-
**Droid Core (GLM-4.7)** and **Droid Core (Kimi K2.5)** are open-source alternatives available in the CLI. They're useful for:
97
+
**Droid Core (GLM-4.7)**, **Droid Core (Kimi K2.5)**, and **Droid Core (MiniMax M2.5)** are open-source alternatives available in the CLI. They're useful for:
94
98
95
99
-**Air-gapped environments** where external API calls aren't allowed
96
100
-**Cost-sensitive projects** needing unlimited local inference
97
101
-**Privacy requirements** where code cannot leave your infrastructure
98
102
-**Experimentation** with open-source model capabilities
99
103
100
-
**Note:** GLM-4.7 does not support image attachments. Kimi K2.5 does support images. For image-based workflows, use Claude, GPT, or Kimi models.
104
+
**Note:** GLM-4.7 and MiniMax M2.5 do not support image attachments. Kimi K2.5 does support images. MiniMax M2.5 is the cheapest model available (0.12× multiplier) and uniquely supports reasoning (Low/Medium/High) among Droid Core models. For image-based workflows, use Claude, GPT, or Kimi models.
101
105
102
106
To use open-source models, you'll need to configure them via BYOK with a local inference server (like Ollama) or a hosted provider. See [BYOK documentation](/cli/configuration/byok) for setup instructions.
![factory-droid[bot]](https://avatars.githubusercontent.com/in/358017?v=4&size=40){kind=avatar}
0 commit comments