torbjorn/my-pal-mcp-server
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: GPT-5.2 support

Browse Source
This commit is contained in:
Fahad
2025-12-11 19:11:50 +00:00
parent 5c3dd75ca6
commit 8b16405f06
33 changed files with 216 additions and 119 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
.env.example
View File

@@ -55,7 +55,7 @@ OPENROUTER_API_KEY=your_openrouter_api_key_here
# Optional: Default model to use
# Options: 'auto' (Claude picks best model), 'pro', 'flash', 'o3', 'o3-mini', 'o4-mini', 'o4-mini-high',
# 'gpt-5.1', 'gpt-5.1-codex', 'gpt-5.1-codex-mini', 'gpt-5', 'gpt-5-mini', 'grok',
# 'gpt-5.2', 'gpt-5.1-codex', 'gpt-5.1-codex-mini', 'gpt-5', 'gpt-5-mini', 'grok',
# 'opus-4.1', 'sonnet-4.1', or any DIAL model if DIAL is configured
# When set to 'auto', Claude will select the best model for each task
# Defaults to 'auto' if not specified
@@ -80,7 +80,8 @@ DEFAULT_THINKING_MODE_THINKDEEP=high
# If you want to disable a provider entirely, don't set its API key
#
# Supported OpenAI models:
# - gpt-5.1 (400K context, 128K output, reasoning tokens, streaming enabled)
# - gpt-5.2 (400K context, 128K output, reasoning tokens, streaming enabled)
# - gpt-5.2-pro (400K context, 272K output, highest reasoning quality, Responses API only)
# - gpt-5.1-codex (400K context, 128K output, coding specialization, Responses API only)
# - gpt-5.1-codex-mini (400K context, 128K output, cost-efficient Codex with streaming)
# - gpt-5 (400K context, 128K output, reasoning tokens)
@@ -126,7 +127,7 @@ DEFAULT_THINKING_MODE_THINKDEEP=high
#
# Examples:
# OPENAI_ALLOWED_MODELS=o3-mini,o4-mini,mini # Only allow mini models (cost control)
# OPENAI_ALLOWED_MODELS=gpt-5.1,gpt-5.1-codex # Pin to GPT-5.1 family
# OPENAI_ALLOWED_MODELS=gpt-5.2,gpt-5.1-codex # Pin to flagship GPT-5 family
# GOOGLE_ALLOWED_MODELS=flash # Only allow Flash (fast responses)
# XAI_ALLOWED_MODELS=grok-3 # Only allow standard GROK (not fast variant)
# OPENAI_ALLOWED_MODELS=o4-mini # Single model standardization

6
README.md
View File

@@ -128,7 +128,7 @@ and review into consideration to aid with its final pre-commit review.
For best results when using [Claude Code](https://claude.ai/code):
- **Sonnet 4.5** - All agentic work and orchestration
- **Gemini 3.0 Pro** OR **GPT-5-Pro** - Deep thinking, additional code reviews, debugging and validations, pre-commit analysis
- **Gemini 3.0 Pro** OR **GPT-5.2-Pro** - Deep thinking, additional code reviews, debugging and validations, pre-commit analysis
</details>
<details>
@@ -137,7 +137,7 @@ For best results when using [Claude Code](https://claude.ai/code):
For best results when using [Codex CLI](https://developers.openai.com/codex/cli):
- **GPT-5 Codex Medium** - All agentic work and orchestration
- **Gemini 3.0 Pro** OR **GPT-5-Pro** - Deep thinking, additional code reviews, debugging and validations, pre-commit analysis
- **Gemini 3.0 Pro** OR **GPT-5.2-Pro** - Deep thinking, additional code reviews, debugging and validations, pre-commit analysis
</details>
## Quick Start (5 minutes)
@@ -208,7 +208,7 @@ PAL activates any provider that has credentials in your `.env`. See `.env.exampl
**Collaboration & Planning** *(Enabled by default)*
- **[`clink`](docs/tools/clink.md)** - Bridge requests to external AI CLIs (Gemini planner, codereviewer, etc.)
- **[`chat`](docs/tools/chat.md)** - Brainstorm ideas, get second opinions, validate approaches. With capable models (GPT-5 Pro, Gemini 3.0 Pro), generates complete code / implementation
- **[`chat`](docs/tools/chat.md)** - Brainstorm ideas, get second opinions, validate approaches. With capable models (GPT-5.2 Pro, Gemini 3.0 Pro), generates complete code / implementation
- **[`thinkdeep`](docs/tools/thinkdeep.md)** - Extended reasoning, edge case analysis, alternative perspectives
- **[`planner`](docs/tools/planner.md)** - Break down complex projects into structured, actionable plans
- **[`consensus`](docs/tools/consensus.md)** - Get expert opinions from multiple AI models with stance steering

2
conf/gemini_models.json
View File

@@ -17,7 +17,7 @@
"max_image_size_mb": "Maximum total size in MB for all images combined (capped at 40MB max for custom models)",
"supports_temperature": "Whether the model accepts temperature parameter in API calls (set to false for O3/O4 reasoning models)",
"temperature_constraint": "Type of temperature constraint: 'fixed' (fixed value), 'range' (continuous range), 'discrete' (specific values), or omit for default range",
"use_openai_response_api": "Set to true when the model must use the /responses endpoint (reasoning models like GPT-5 Pro). Leave false/omit for standard chat completions.",
"use_openai_response_api": "Set to true when the model must use the /responses endpoint (reasoning models like GPT-5.2 Pro). Leave false/omit for standard chat completions.",
"default_reasoning_effort": "Default reasoning effort level for models that support it (e.g., 'low', 'medium', 'high'). Omit if not applicable.",
"description": "Human-readable description of the model",
"intelligence_score": "1-20 human rating used as the primary signal for auto-mode model ordering",

19
conf/openai_models.json
View File

@@ -17,7 +17,7 @@
"max_image_size_mb": "Maximum total size in MB for all images combined (capped at 40MB max for custom models)",
"supports_temperature": "Whether the model accepts temperature parameter in API calls (set to false for O3/O4 reasoning models)",
"temperature_constraint": "Type of temperature constraint: 'fixed' (fixed value), 'range' (continuous range), 'discrete' (specific values), or omit for default range",
"use_openai_response_api": "Set to true when the model must use the /responses endpoint (reasoning models like GPT-5 Pro). Leave false/omit for standard chat completions.",
"use_openai_response_api": "Set to true when the model must use the /responses endpoint (reasoning models like GPT-5.2 Pro). Leave false/omit for standard chat completions.",
"default_reasoning_effort": "Default reasoning effort level for models that support it (e.g., 'low', 'medium', 'high'). Omit if not applicable.",
"description": "Human-readable description of the model",
"intelligence_score": "1-20 human rating used as the primary signal for auto-mode model ordering",
@@ -47,14 +47,16 @@
"temperature_constraint": "fixed"
},
{
"model_name": "gpt-5-pro",
"friendly_name": "OpenAI (GPT-5 Pro)",
"model_name": "gpt-5.2-pro",
"friendly_name": "OpenAI (GPT-5.2 Pro)",
"aliases": [
"gpt5.2-pro",
"gpt5.2pro",
"gpt5pro",
"gpt5-pro"
],
"intelligence_score": 18,
"description": "GPT-5 Pro (400K context, 272K output) - Very advanced, reasoning model",
"description": "GPT-5.2 Pro (400K context, 272K output) - Very advanced, reasoning model",
"context_window": 400000,
"max_output_tokens": 272000,
"supports_extended_thinking": true,
@@ -234,15 +236,18 @@
"use_openai_response_api": true
},
{
"model_name": "gpt-5.1",
"friendly_name": "OpenAI (GPT-5.1)",
"model_name": "gpt-5.2",
"friendly_name": "OpenAI (GPT-5.2)",
"aliases": [
"gpt5.2",
"gpt-5.2",
"5.2",
"gpt5.1",
"gpt-5.1",
"5.1"
],
"intelligence_score": 18,
"description": "GPT-5.1 (400K context, 128K output) - Flagship reasoning model with configurable thinking effort and vision support.",
"description": "GPT-5.2 (400K context, 128K output) - Flagship reasoning model with configurable thinking effort and vision support.",
"context_window": 400000,
"max_output_tokens": 128000,
"supports_extended_thinking": true,

15
conf/openrouter_models.json
View File

@@ -16,7 +16,7 @@
"max_image_size_mb": "Maximum total size in MB for all images combined (capped at 40MB max for custom models)",
"supports_temperature": "Whether the model accepts temperature parameter in API calls (set to false for O3/O4 reasoning models)",
"temperature_constraint": "Type of temperature constraint: 'fixed' (fixed value), 'range' (continuous range), 'discrete' (specific values), or omit for default range",
"use_openai_response_api": "Set to true when the model must use the /responses endpoint (reasoning models like GPT-5 Pro). Leave false/omit for standard chat completions.",
"use_openai_response_api": "Set to true when the model must use the /responses endpoint (reasoning models like GPT-5.2 Pro). Leave false/omit for standard chat completions.",
"default_reasoning_effort": "Default reasoning effort level for models that support it (e.g., 'low', 'medium', 'high'). Omit if not applicable.",
"description": "Human-readable description of the model",
"intelligence_score": "1-20 human rating used as the primary signal for auto-mode model ordering",
@@ -315,8 +315,10 @@
"intelligence_score": 16
},
{
"model_name": "openai/gpt-5-pro",
"model_name": "openai/gpt-5.2-pro",
"aliases": [
"gpt5.2-pro",
"gpt5.2pro",
"gpt5pro"
],
"context_window": 400000,
@@ -331,7 +333,7 @@
"use_openai_response_api": true,
"default_reasoning_effort": "high",
"allow_code_generation": true,
"description": "GPT-5 Pro - Advanced reasoning model with highest quality responses (text+image input, text output only)",
"description": "GPT-5.2 Pro - Advanced reasoning model with highest quality responses (text+image input, text output only)",
"intelligence_score": 18
},
{
@@ -385,8 +387,11 @@
"intelligence_score": 8
},
{
"model_name": "openai/gpt-5.1",
"model_name": "openai/gpt-5.2",
"aliases": [
"gpt5.2",
"gpt-5.2",
"5.2",
"gpt5.1",
"gpt-5.1",
"5.1"
@@ -402,7 +407,7 @@
"temperature_constraint": "fixed",
"default_reasoning_effort": "medium",
"allow_code_generation": true,
"description": "GPT-5.1 (400K context, 128K output) - Flagship reasoning model with configurable thinking effort and vision support",
"description": "GPT-5.2 (400K context, 128K output) - Flagship reasoning model with configurable thinking effort and vision support",
"intelligence_score": 18
},
{

2
conf/xai_models.json
View File

@@ -17,7 +17,7 @@
"max_image_size_mb": "Maximum total size in MB for all images combined (capped at 40MB max for custom models)",
"supports_temperature": "Whether the model accepts temperature parameter in API calls (set to false for O3/O4 reasoning models)",
"temperature_constraint": "Type of temperature constraint: 'fixed' (fixed value), 'range' (continuous range), 'discrete' (specific values), or omit for default range",
"use_openai_response_api": "Set to true when the model must use the /responses endpoint (reasoning models like GPT-5 Pro). Leave false/omit for standard chat completions.",
"use_openai_response_api": "Set to true when the model must use the /responses endpoint (reasoning models like GPT-5.2 Pro). Leave false/omit for standard chat completions.",
"default_reasoning_effort": "Default reasoning effort level for models that support it (e.g., 'low', 'medium', 'high'). Omit if not applicable.",
"description": "Human-readable description of the model",
"intelligence_score": "1-20 human rating used as the primary signal for auto-mode model ordering"

18
docs/advanced-usage.md
View File

@@ -41,7 +41,7 @@ Regardless of your default configuration, you can specify models per request:
| **`o3-mini`** | OpenAI | 200K tokens | Balanced speed/quality | Moderate complexity tasks |
| **`o4-mini`** | OpenAI | 200K tokens | Latest reasoning model | Optimized for shorter contexts |
| **`gpt4.1`** | OpenAI | 1M tokens | Latest GPT-4 with extended context | Large codebase analysis, comprehensive reviews |
| **`gpt5.1`** (GPT-5.1) | OpenAI | 400K tokens | Flagship reasoning model with configurable thinking effort | Complex problems, balanced agent/coding flows |
| **`gpt5.2`** (GPT-5.2) | OpenAI | 400K tokens | Flagship reasoning model with configurable thinking effort | Complex problems, balanced agent/coding flows |
| **`gpt5.1-codex`** (GPT-5.1 Codex) | OpenAI | 400K tokens | Agentic coding specialization (Responses API) | Advanced coding tasks, structured code generation |
| **`gpt5.1-codex-mini`** (GPT-5.1 Codex mini) | OpenAI | 400K tokens | Cost-efficient Codex variant with streaming | Balanced coding tasks, cost-conscious development |
| **`gpt5`** (GPT-5) | OpenAI | 400K tokens | Advanced model with reasoning support | Complex problems requiring advanced reasoning |
@@ -64,8 +64,8 @@ cloud models (expensive/powerful) AND local models (free/private) in the same co
- **Flash Lite 2.0**: Text-only lightweight model (no thinking support)
- **O3/O4 Models**: Excellent reasoning, systematic analysis, 200K context
- **GPT-4.1**: Extended context window (1M tokens), general capabilities
- **GPT-5.1 Series**: Latest flagship reasoning models, 400K context
- **GPT-5.1**: Flagship model with configurable thinking effort and vision
- **GPT-5.2 Series**: Latest flagship reasoning models, 400K context
- **GPT-5.2**: Flagship model with configurable thinking effort and vision
- **GPT-5.1 Codex**: Agentic coding specialization (Responses API, non-streaming)
- **GPT-5.1 Codex mini**: Cost-efficient Codex variant with streaming support
- **GPT-5 Series**: Advanced reasoning models, 400K context
@@ -168,7 +168,7 @@ All tools that work with files support **both individual files and entire direct
**`analyze`** - Analyze files or directories
- `files`: List of file paths or directories (required)
- `question`: What to analyze (required)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `analysis_type`: architecture|performance|security|quality|general
- `output_format`: summary|detailed|actionable
- `thinking_mode`: minimal|low|medium|high|max (default: medium, Gemini only)
@@ -183,7 +183,7 @@ All tools that work with files support **both individual files and entire direct
**`codereview`** - Review code files or directories
- `files`: List of file paths or directories (required)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `review_type`: full|security|performance|quick
- `focus_on`: Specific aspects to focus on
- `standards`: Coding standards to enforce
@@ -199,7 +199,7 @@ All tools that work with files support **both individual files and entire direct
**`debug`** - Debug with file context
- `error_description`: Description of the issue (required)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `error_context`: Stack trace or logs
- `files`: Files or directories related to the issue
- `runtime_info`: Environment details
@@ -215,7 +215,7 @@ All tools that work with files support **both individual files and entire direct
**`thinkdeep`** - Extended analysis with file context
- `current_analysis`: Your current thinking (required)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `problem_context`: Additional context
- `focus_areas`: Specific aspects to focus on
- `files`: Files or directories for context
@@ -231,7 +231,7 @@ All tools that work with files support **both individual files and entire direct
**`testgen`** - Comprehensive test generation with edge case coverage
- `files`: Code files or directories to generate tests for (required)
- `prompt`: Description of what to test, testing objectives, and scope (required)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `test_examples`: Optional existing test files as style/pattern reference
- `thinking_mode`: minimal|low|medium|high|max (default: medium, Gemini only)
@@ -246,7 +246,7 @@ All tools that work with files support **both individual files and entire direct
- `files`: Code files or directories to analyze for refactoring opportunities (required)
- `prompt`: Description of refactoring goals, context, and specific areas of focus (required)
- `refactor_type`: codesmells|decompose|modernize|organization (required)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `focus_areas`: Specific areas to focus on (e.g., 'performance', 'readability', 'maintainability', 'security')
- `style_guide_examples`: Optional existing code files to use as style/pattern reference
- `thinking_mode`: minimal|low|medium|high|max (default: medium, Gemini only)

16
docs/configuration.md
View File

@@ -63,7 +63,7 @@ CUSTOM_MODEL_NAME=llama3.2 # Default model
**Default Model Selection:**
```env
# Options: 'auto', 'pro', 'flash', 'gpt5.1', 'gpt5.1-codex', 'gpt5.1-codex-mini', 'o3', 'o3-mini', 'o4-mini', etc.
# Options: 'auto', 'pro', 'flash', 'gpt5.2', 'gpt5.1-codex', 'gpt5.1-codex-mini', 'o3', 'o3-mini', 'o4-mini', etc.
DEFAULT_MODEL=auto # Claude picks best model for each task (recommended)
```
@@ -81,13 +81,13 @@ DEFAULT_MODEL=auto # Claude picks best model for each task (recommended)
| Provider | Canonical Models | Notable Aliases |
|----------|-----------------|-----------------|
| OpenAI | `gpt-5.1`, `gpt-5.1-codex`, `gpt-5.1-codex-mini`, `gpt-5`, `gpt-5-pro`, `gpt-5-mini`, `gpt-5-nano`, `gpt-5-codex`, `gpt-4.1`, `o3`, `o3-mini`, `o3-pro`, `o4-mini` | `gpt5.1`, `gpt-5.1`, `5.1`, `gpt5.1-codex`, `codex-5.1`, `codex-mini`, `gpt5`, `gpt5pro`, `mini`, `nano`, `codex`, `o3mini`, `o3pro`, `o4mini` |
| OpenAI | `gpt-5.2`, `gpt-5.1-codex`, `gpt-5.1-codex-mini`, `gpt-5`, `gpt-5.2-pro`, `gpt-5-mini`, `gpt-5-nano`, `gpt-5-codex`, `gpt-4.1`, `o3`, `o3-mini`, `o3-pro`, `o4-mini` | `gpt5.2`, `gpt-5.2`, `5.2`, `gpt5.1-codex`, `codex-5.1`, `codex-mini`, `gpt5`, `gpt5pro`, `mini`, `nano`, `codex`, `o3mini`, `o3pro`, `o4mini` |
| Gemini | `gemini-2.5-pro`, `gemini-2.5-flash`, `gemini-2.0-flash`, `gemini-2.0-flash-lite` | `pro`, `gemini-pro`, `flash`, `flash-2.0`, `flashlite` |
| X.AI | `grok-4`, `grok-3`, `grok-3-fast` | `grok`, `grok4`, `grok3`, `grok3fast`, `grokfast` |
| OpenRouter | See `conf/openrouter_models.json` for the continually evolving catalogue | e.g., `opus`, `sonnet`, `flash`, `pro`, `mistral` |
| Custom | User-managed entries such as `llama3.2` | Define your own aliases per entry |
Latest OpenAI entries (`gpt-5.1`, `gpt-5.1-codex`, `gpt-5.1-codex-mini`) mirror the official model cards released on November 13, 2025: all three expose 400K-token contexts with 128K-token outputs, reasoning-token support, and multimodal inputs. `gpt-5.1-codex` is Responses-only with streaming disabled, while the base `gpt-5.1` and Codex mini support streaming along with full code-generation flags. Update your manifests if you run custom deployments so these capability bits stay accurate.
Latest OpenAI entries (`gpt-5.2`, `gpt-5.1-codex`, `gpt-5.1-codex-mini`, `gpt-5.2-pro`) expose 400K-token contexts with large outputs, reasoning-token support, and multimodal inputs. `gpt-5.1-codex` and `gpt-5.2-pro` are Responses-only with streaming disabled, while the base `gpt-5.2` and Codex mini support streaming along with full code-generation flags. Update your manifests if you run custom deployments so these capability bits stay accurate.
> **Tip:** Copy the JSON file you need, customise it, and point the corresponding `*_MODELS_CONFIG_PATH` environment variable to your version. This lets you enable or disable capabilities (JSON mode, function calling, temperature support, code generation) without editing Python.
@@ -107,7 +107,7 @@ The `allow_code_generation` capability enables models to generate complete, prod
**When to Enable:**
- **Enable for**: Models MORE capable than your primary CLI's model (e.g., GPT-5.1 Codex, GPT-5 Pro, GPT-5.1 when using Claude Code with Sonnet 4.5)
- **Enable for**: Models MORE capable than your primary CLI's model (e.g., GPT-5.1 Codex, GPT-5.2 Pro, GPT-5.2 when using Claude Code with Sonnet 4.5)
- **Purpose**: Get complete implementations from a more powerful reasoning model that your primary CLI can then review and apply
- **Use case**: Large-scale implementations, major refactoring, complete module creation
@@ -132,7 +132,7 @@ The `allow_code_generation` capability enables models to generate complete, prod
...
},
{
"model_name": "gpt-5-pro",
"model_name": "gpt-5.2-pro",
"allow_code_generation": true,
"intelligence_score": 19,
...
@@ -142,8 +142,8 @@ The `allow_code_generation` capability enables models to generate complete, prod
```
**Typical Workflow:**
1. You ask your AI agent to implement a complex new feature using `chat` with a higher-reasoning model such as **gpt-5-pro**
2. GPT-5-Pro generates structured implementation and shares the complete implementation with PAL
1. You ask your AI agent to implement a complex new feature using `chat` with a higher-reasoning model such as **gpt-5.2-pro**
2. GPT-5.2-Pro generates structured implementation and shares the complete implementation with PAL
3. PAL saves the code to `pal_generated.code` and asks AI agent to implement the plan
4. AI agent continues from the previous context, reads the file, applies the implementation
@@ -198,7 +198,7 @@ OPENAI_ALLOWED_MODELS=o4-mini
GOOGLE_ALLOWED_MODELS=flash
# High-performance setup
OPENAI_ALLOWED_MODELS=gpt-5.1-codex,gpt-5.1
OPENAI_ALLOWED_MODELS=gpt-5.1-codex,gpt-5.2
GOOGLE_ALLOWED_MODELS=pro
# Single model standardization

10
docs/custom_models.md
View File

@@ -61,7 +61,7 @@ The curated defaults in `conf/openrouter_models.json` include popular entries su
| `llama3` | `meta-llama/llama-3-70b` | Large open-weight text model |
| `deepseek-r1` | `deepseek/deepseek-r1-0528` | DeepSeek reasoning model |
| `perplexity` | `perplexity/llama-3-sonar-large-32k-online` | Search-augmented model |
| `gpt5.1`, `gpt-5.1`, `5.1` | `openai/gpt-5.1` | Flagship GPT-5.1 with reasoning and vision |
| `gpt5.2`, `gpt-5.2`, `5.2` | `openai/gpt-5.2` | Flagship GPT-5.2 with reasoning and vision |
| `gpt5.1-codex`, `codex-5.1` | `openai/gpt-5.1-codex` | Agentic coding specialization (Responses API) |
| `codex-mini`, `gpt5.1-codex-mini` | `openai/gpt-5.1-codex-mini` | Cost-efficient Codex variant with streaming |
@@ -77,21 +77,21 @@ View the baseline OpenRouter catalogue in [`conf/openrouter_models.json`](conf/o
Native catalogues (`conf/openai_models.json`, `conf/gemini_models.json`, `conf/xai_models.json`, `conf/dial_models.json`) follow the same schema. Updating those files lets you:
- Expose new aliases (e.g., map `enterprise-pro` to `gpt-5-pro`)
- Expose new aliases (e.g., map `enterprise-pro` to `gpt-5.2-pro`)
- Advertise support for JSON mode or vision if the upstream provider adds it
- Adjust token limits when providers increase context windows
### Latest OpenAI releases
OpenAI's November 13, 2025 drop introduced `gpt-5.1`, `gpt-5.1-codex`, and `gpt-5.1-codex-mini`, all of which now ship in `conf/openai_models.json`:
OpenAI's November 13, 2025 drop introduced `gpt-5.1-codex` and `gpt-5.1-codex-mini`, while the flagship base model is now `gpt-5.2`. All of these ship in `conf/openai_models.json`:
| Model | Highlights | Notes |
|-------|------------|-------|
| `gpt-5.1` | 400K context, 128K output, multimodal IO, configurable reasoning effort | Streaming enabled; use for balanced agent/coding flows |
| `gpt-5.2` | 400K context, 128K output, multimodal IO, configurable reasoning effort | Streaming enabled; use for balanced agent/coding flows |
| `gpt-5.1-codex` | Responses-only agentic coding version of GPT-5.1 | Streaming disabled; `use_openai_response_api=true`; `allow_code_generation=true` |
| `gpt-5.1-codex-mini` | Cost-efficient Codex variant | Streaming enabled, retains 400K context and code-generation flag |
These entries include pricing-friendly aliases (`gpt5.1`, `codex-5.1`, `codex-mini`) plus updated capability flags (`supports_extended_thinking`, `allow_code_generation`). Copy the manifest if you operate custom deployment names so downstream providers inherit the same metadata.
These entries include pricing-friendly aliases (`gpt5.2`, `codex-5.1`, `codex-mini`) plus updated capability flags (`supports_extended_thinking`, `allow_code_generation`). Copy the manifest if you operate custom deployment names so downstream providers inherit the same metadata.
Because providers load the manifests on import, you can tweak capabilities without touching Python. Restart the server after editing the JSON files so changes are picked up.

6
docs/getting-started.md
View File

@@ -29,7 +29,7 @@ You need at least one API key. Choose based on your needs:
**OpenAI:**
- Visit [OpenAI Platform](https://platform.openai.com/api-keys)
- Generate an API key for GPT-5.1, GPT-5.1-Codex, GPT-5, O3 access
- Generate an API key for GPT-5.2, GPT-5.1-Codex, GPT-5, O3 access
**X.AI (Grok):**
- Visit [X.AI Console](https://console.x.ai/)
@@ -287,7 +287,7 @@ Add your API keys (at least one required):
```env
# Choose your providers (at least one required)
GEMINI_API_KEY=your-gemini-api-key-here # For Gemini models
OPENAI_API_KEY=your-openai-api-key-here # For GPT-5.1, GPT-5.1-Codex, O3
OPENAI_API_KEY=your-openai-api-key-here # For GPT-5.2, GPT-5.1-Codex, O3
XAI_API_KEY=your-xai-api-key-here # For Grok models
OPENROUTER_API_KEY=your-openrouter-key # For multiple models
@@ -514,7 +514,7 @@ DEFAULT_MODEL=auto
GEMINI_API_KEY=your-key
OPENAI_API_KEY=your-key
GOOGLE_ALLOWED_MODELS=pro
OPENAI_ALLOWED_MODELS=gpt-5.1-codex,gpt-5.1
OPENAI_ALLOWED_MODELS=gpt-5.1-codex,gpt-5.2
```
### Local-First Setup

2
docs/model_ranking.md
View File

@@ -39,7 +39,7 @@ A straightforward rubric that mirrors typical provider tiers:
| Intelligence | Guidance |
|--------------|-------------------------------------------------------------------------------------------|
| 18–19 | Frontier reasoning models (Gemini 3.0 Pro, Gemini 2.5 Pro, GPT‑5.1 Codex, GPT‑5.1, GPT‑5) |
| 18–19 | Frontier reasoning models (Gemini 3.0 Pro, Gemini 2.5 Pro, GPT‑5.1 Codex, GPT‑5.2 Pro, GPT‑5.2, GPT‑5) |
| 15–17 | Strong general models with large context (O3 Pro, DeepSeek R1) |
| 12–14 | Balanced assistants (Claude Opus/Sonnet, Mistral Large) |
| 9–11 | Fast distillations (Gemini Flash, GPT-5 Mini, Mistral medium) |

2
docs/tools/analyze.md
View File

@@ -64,7 +64,7 @@ This workflow ensures methodical analysis before expert insights, resulting in d
**Initial Configuration (used in step 1):**
- `prompt`: What to analyze or look for (required)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `analysis_type`: architecture|performance|security|quality|general (default: general)
- `output_format`: summary|detailed|actionable (default: detailed)
- `temperature`: Temperature for analysis (0-1, default 0.2)

10
docs/tools/chat.md
View File

@@ -39,7 +39,7 @@ word verdict in the end.
- **Collaborative thinking partner** for your analysis and planning
- **Get second opinions** on your designs and approaches
- **Brainstorm solutions** and explore alternatives together
- **Structured code generation**: When using GPT-5.1 or Gemini 3.0 / 2.5 Pro, get complete, production-ready implementations saved to `pal_generated.code` for your CLI to review and apply
- **Structured code generation**: When using GPT-5.2 or Gemini 3.0 / 2.5 Pro, get complete, production-ready implementations saved to `pal_generated.code` for your CLI to review and apply
- **Validate your checklists** and implementation plans
- **General development questions** and explanations
- **Technology comparisons** and best practices
@@ -52,7 +52,7 @@ word verdict in the end.
## Tool Parameters
- `prompt`: Your question or discussion topic (required)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `absolute_file_paths`: Optional absolute file or directory paths for additional context
- `images`: Optional images for visual context (absolute paths)
- `working_directory_absolute_path`: **Required** - Absolute path to an existing directory where generated code artifacts will be saved
@@ -62,11 +62,11 @@ word verdict in the end.
## Structured Code Generation
When using advanced reasoning models like **GPT-5 Pro** or **Gemini 3.0 Pro**, the chat tool can generate complete, production-ready code implementations in a structured format.
When using advanced reasoning models like **GPT-5.2 Pro** or **Gemini 3.0 Pro**, the chat tool can generate complete, production-ready code implementations in a structured format.
### How It Works
1. You ask your AI agent to implement a complex new feature using `chat` with a higher-reasoning model such as **GPT-5 Pro** or **Gemini 3.0 Pro**
1. You ask your AI agent to implement a complex new feature using `chat` with a higher-reasoning model such as **GPT-5.2 Pro** or **Gemini 3.0 Pro**
2. The model generates structured implementation and shares the complete implementation with PAL
3. PAL saves the code to `pal_generated.code` and asks AI agent to implement the plan
4. AI agent continues from the previous context, reads the file, applies the implementation
@@ -85,7 +85,7 @@ For minor changes (small tweaks, bug fixes, algorithm improvements), the model r
### Example Usage
```
chat with gpt-5-pro and ask it to make me a standalone, classic version of the
chat with gpt-5.2-pro and ask it to make me a standalone, classic version of the
Pacman game using pygame that I can run from the commandline. Give me a single
script to execute in the end with any / all dependencies setup for me.
Do everything using pygame, we have no external resources / images / audio at

2
docs/tools/codereview.md
View File

@@ -79,7 +79,7 @@ The above prompt will simultaneously run two separate `codereview` tools with tw
**Initial Review Configuration (used in step 1):**
- `prompt`: User's summary of what the code does, expected behavior, constraints, and review objectives (required)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `review_type`: full|security|performance|quick (default: full)
- `focus_on`: Specific aspects to focus on (e.g., "security vulnerabilities", "performance bottlenecks")
- `standards`: Coding standards to enforce (e.g., "PEP8", "ESLint", "Google Style Guide")

2
docs/tools/debug.md
View File

@@ -72,7 +72,7 @@ This structured approach ensures Claude performs methodical groundwork before ex
- `images`: Visual debugging materials (error screenshots, logs, etc.)
**Model Selection:**
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `thinking_mode`: minimal|low|medium|high|max (default: medium, Gemini only)
- `use_assistant_model`: Whether to use expert analysis phase (default: true, set to false to use Claude only)

2
docs/tools/precommit.md
View File

@@ -140,7 +140,7 @@ Use pal and perform a thorough precommit ensuring there aren't any new regressio
**Initial Configuration (used in step 1):**
- `path`: Starting directory to search for repos (REQUIRED for step 1, must be absolute path)
- `prompt`: The original user request description for the changes (required for context)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `compare_to`: Compare against a branch/tag instead of local changes (optional)
- `severity_filter`: critical|high|medium|low|all (default: all)
- `include_staged`: Include staged changes in the review (default: true)

2
docs/tools/refactor.md
View File

@@ -102,7 +102,7 @@ This results in Claude first performing its own expert analysis, encouraging it
**Initial Configuration (used in step 1):**
- `prompt`: Description of refactoring goals, context, and specific areas of focus (required)
- `refactor_type`: codesmells|decompose|modernize|organization (default: codesmells)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `focus_areas`: Specific areas to focus on (e.g., 'performance', 'readability', 'maintainability', 'security')
- `style_guide_examples`: Optional existing code files to use as style/pattern reference (absolute paths)
- `thinking_mode`: minimal|low|medium|high|max (default: medium, Gemini only)

2
docs/tools/secaudit.md
View File

@@ -85,7 +85,7 @@ security remediation plan using planner
- `images`: Architecture diagrams, security documentation, or visual references
**Initial Security Configuration (used in step 1):**
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `security_scope`: Application context, technology stack, and security boundary definition (required)
- `threat_level`: low|medium|high|critical (default: medium) - determines assessment depth and urgency
- `compliance_requirements`: List of compliance frameworks to assess against (e.g., ["PCI DSS", "SOC2"])

2
docs/tools/testgen.md
View File

@@ -69,7 +69,7 @@ Test generation excels with extended reasoning models like Gemini Pro or O3, whi
**Initial Configuration (used in step 1):**
- `prompt`: Description of what to test, testing objectives, and specific scope/focus areas (required)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `test_examples`: Optional existing test files or directories to use as style/pattern reference (absolute paths)
- `thinking_mode`: minimal|low|medium|high|max (default: medium, Gemini only)
- `use_assistant_model`: Whether to use expert test generation phase (default: true, set to false to use Claude only)

2
docs/tools/thinkdeep.md
View File

@@ -30,7 +30,7 @@ with the best architecture for my project
## Tool Parameters
- `prompt`: Your current thinking/analysis to extend and validate (required)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.1|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5.2|gpt5.1-codex|gpt5.1-codex-mini|gpt5|gpt5-mini|gpt5-nano (default: server default)
- `problem_context`: Additional context about the problem or goal
- `focus_areas`: Specific aspects to focus on (architecture, performance, security, etc.)
- `files`: Optional file paths or directories for additional context (absolute paths)

14
providers/openai.py
View File

@@ -119,9 +119,9 @@ class OpenAIModelProvider(RegistryBackedProviderMixin, OpenAICompatibleProvider)
preferred = find_first(
[
"gpt-5.1-codex",
"gpt-5.1",
"gpt-5.2",
"gpt-5-codex",
"gpt-5-pro",
"gpt-5.2-pro",
"o3-pro",
"gpt-5",
"o3",
@@ -131,10 +131,10 @@ class OpenAIModelProvider(RegistryBackedProviderMixin, OpenAICompatibleProvider)
elif category == ToolModelCategory.FAST_RESPONSE:
# Prefer fast, cost-efficient models
# GPT-5.1 models for speed, GPT-5.1-Codex after (premium pricing but cached)
# GPT-5.2 models for speed, GPT-5.1-Codex after (premium pricing but cached)
preferred = find_first(
[
"gpt-5.1",
"gpt-5.2",
"gpt-5.1-codex-mini",
"gpt-5",
"gpt-5-mini",
@@ -147,14 +147,14 @@ class OpenAIModelProvider(RegistryBackedProviderMixin, OpenAICompatibleProvider)
else: # BALANCED or default
# Prefer balanced performance/cost models
# Include GPT-5.1 family for latest capabilities
# Include GPT-5.2 family for latest capabilities
preferred = find_first(
[
"gpt-5.1",
"gpt-5.2",
"gpt-5.1-codex",
"gpt-5",
"gpt-5-codex",
"gpt-5-pro",
"gpt-5.2-pro",
"gpt-5-mini",
"o4-mini",
"o3-mini",

10
tests/CASSETTE_MAINTENANCE.md
View File

@@ -228,10 +228,10 @@ Some integration tests maintain cassettes for multiple model variants to ensure
### Consensus Tool Cassettes
The `test_consensus_integration.py` test uses parameterized fixtures to test both `gpt-5` and `gpt-5.1` models:
The `test_consensus_integration.py` test uses parameterized fixtures to test both `gpt-5` and `gpt-5.2` models:
- `tests/openai_cassettes/consensus_step1_gpt5_for.json` - Cassette for gpt-5 model
- `tests/openai_cassettes/consensus_step1_gpt51_for.json` - Cassette for gpt-5.1 model
- `tests/openai_cassettes/consensus_step1_gpt52_for.json` - Cassette for gpt-5.2 model
**When updating consensus cassettes:**
@@ -249,9 +249,9 @@ rm tests/openai_cassettes/consensus_step1_gpt5_for.json
# Run the test with real API key (it will record for gpt-5)
OPENAI_API_KEY="your-real-key" python -m pytest tests/test_consensus_integration.py::test_consensus_multi_model_consultations[gpt-5] -v
# Or for gpt-5.1
rm tests/openai_cassettes/consensus_step1_gpt51_for.json
OPENAI_API_KEY="your-real-key" python -m pytest tests/test_consensus_integration.py::test_consensus_multi_model_consultations[gpt-5.1] -v
# Or for gpt-5.2
rm tests/openai_cassettes/consensus_step1_gpt52_for.json
OPENAI_API_KEY="your-real-key" python -m pytest tests/test_consensus_integration.py::test_consensus_multi_model_consultations[gpt-5.2] -v
```
This dual-coverage approach ensures that both model families continue to work correctly as the codebase evolves.

28
tests/gemini_cassettes/consensus/step2_gemini25_flash_against/mldev.json
View File

File diff suppressed because one or more lines are too long

82
tests/openai_cassettes/consensus_step1_gpt52_for.json Normal file
View File

File diff suppressed because one or more lines are too long

4
tests/test_auto_mode_comprehensive.py
View File

@@ -95,8 +95,8 @@ class TestAutoModeComprehensive:
},
{
"EXTENDED_REASONING": "gpt-5.1-codex", # GPT-5.1 Codex prioritized for coding tasks
"FAST_RESPONSE": "gpt-5.1", # Prefer gpt-5.1 for speed
"BALANCED": "gpt-5.1", # Prefer gpt-5.1 for balanced
"FAST_RESPONSE": "gpt-5.2", # Prefer gpt-5.2 for speed
"BALANCED": "gpt-5.2", # Prefer gpt-5.2 for balanced
},
),
# Only X.AI API available

8
tests/test_auto_mode_model_listing.py
View File

@@ -83,7 +83,7 @@ def test_error_listing_respects_env_restrictions(monkeypatch, reset_registry):
pass
monkeypatch.setenv("GOOGLE_ALLOWED_MODELS", "gemini-2.5-pro")
monkeypatch.setenv("OPENAI_ALLOWED_MODELS", "gpt-5.1")
monkeypatch.setenv("OPENAI_ALLOWED_MODELS", "gpt-5.2")
monkeypatch.setenv("OPENROUTER_ALLOWED_MODELS", "gpt5nano")
monkeypatch.setenv("XAI_ALLOWED_MODELS", "")
@@ -104,7 +104,7 @@ def test_error_listing_respects_env_restrictions(monkeypatch, reset_registry):
("OPENAI_API_KEY", "test-openai"),
("OPENROUTER_API_KEY", "test-openrouter"),
("GOOGLE_ALLOWED_MODELS", "gemini-2.5-pro"),
("OPENAI_ALLOWED_MODELS", "gpt-5.1"),
("OPENAI_ALLOWED_MODELS", "gpt-5.2"),
("OPENROUTER_ALLOWED_MODELS", "gpt5nano"),
("XAI_ALLOWED_MODELS", ""),
):
@@ -139,7 +139,7 @@ def test_error_listing_respects_env_restrictions(monkeypatch, reset_registry):
assert payload["status"] == "error"
available_models = _extract_available_models(payload["content"])
assert set(available_models) == {"gemini-2.5-pro", "gpt-5.1", "gpt5nano", "openai/gpt-5-nano"}
assert set(available_models) == {"gemini-2.5-pro", "gpt-5.2", "gpt5nano", "openai/gpt-5-nano"}
@pytest.mark.no_mock_provider
@@ -225,6 +225,6 @@ def test_error_listing_without_restrictions_shows_full_catalog(monkeypatch, rese
available_models = _extract_available_models(payload["content"])
assert "gemini-2.5-pro" in available_models
assert any(model in available_models for model in {"gpt-5.1", "gpt-5"})
assert any(model in available_models for model in {"gpt-5.2", "gpt-5"})
assert "grok-4" in available_models
assert len(available_models) >= 5

4
tests/test_auto_mode_provider_selection.py
View File

@@ -99,8 +99,8 @@ class TestAutoModeProviderSelection:
# Should select appropriate OpenAI models based on new preference order
assert extended_reasoning == "gpt-5.1-codex" # GPT-5.1 Codex prioritized for extended reasoning
assert fast_response == "gpt-5.1" # gpt-5.1 comes first in fast response preference
assert balanced == "gpt-5.1" # gpt-5.1 for balanced
assert fast_response == "gpt-5.2" # gpt-5.2 comes first in fast response preference
assert balanced == "gpt-5.2" # gpt-5.2 for balanced
finally:
# Restore original environment

6
tests/test_consensus_integration.py
View File

@@ -20,7 +20,7 @@ CASSETTE_DIR.mkdir(exist_ok=True)
# Mapping of OpenAI model names to their cassette files
CONSENSUS_CASSETTES = {
"gpt-5": CASSETTE_DIR / "consensus_step1_gpt5_for.json",
"gpt-5.1": CASSETTE_DIR / "consensus_step1_gpt51_for.json",
"gpt-5.2": CASSETTE_DIR / "consensus_step1_gpt52_for.json",
}
GEMINI_REPLAY_DIR = Path(__file__).parent / "gemini_cassettes"
@@ -32,11 +32,11 @@ GEMINI_REPLAY_PATH = GEMINI_REPLAY_DIR / "consensus" / "step2_gemini25_flash_aga
@pytest.mark.integration
@pytest.mark.asyncio
@pytest.mark.no_mock_provider
@pytest.mark.parametrize("openai_model", ["gpt-5", "gpt-5.1"])
@pytest.mark.parametrize("openai_model", ["gpt-5", "gpt-5.2"])
async def test_consensus_multi_model_consultations(monkeypatch, openai_model):
"""Exercise ConsensusTool against OpenAI model (supporting) and gemini-2.5-flash (critical).
Tests both gpt-5 and gpt-5.1 to ensure regression coverage for both model families.
Tests both gpt-5 and gpt-5.2 to ensure regression coverage for both model families.
"""
# Get the cassette path for this model

8
tests/test_intelligent_fallback.py
View File

@@ -37,14 +37,14 @@ class TestIntelligentFallback:
@patch.dict(os.environ, {"OPENAI_API_KEY": "sk-test-key", "GEMINI_API_KEY": ""}, clear=False)
def test_prefers_openai_o3_mini_when_available(self):
"""Test that gpt-5.1 is preferred when OpenAI API key is available (based on new preference order)"""
"""Test that gpt-5.2 is preferred when OpenAI API key is available (based on new preference order)"""
# Register only OpenAI provider for this test
from providers.openai import OpenAIModelProvider
ModelProviderRegistry.register_provider(ProviderType.OPENAI, OpenAIModelProvider)
fallback_model = ModelProviderRegistry.get_preferred_fallback_model()
assert fallback_model == "gpt-5.1" # Based on new preference order: gpt-5.1 before o4-mini
assert fallback_model == "gpt-5.2" # Based on new preference order: gpt-5.2 before o4-mini
@patch.dict(os.environ, {"OPENAI_API_KEY": "", "GEMINI_API_KEY": "test-gemini-key"}, clear=False)
def test_prefers_gemini_flash_when_openai_unavailable(self):
@@ -147,8 +147,8 @@ class TestIntelligentFallback:
history, tokens = build_conversation_history(context, model_context=None)
# Verify that ModelContext was called with gpt-5.1 (the intelligent fallback based on new preference order)
mock_context_class.assert_called_once_with("gpt-5.1")
# Verify that ModelContext was called with gpt-5.2 (the intelligent fallback based on new preference order)
mock_context_class.assert_called_once_with("gpt-5.2")
def test_auto_mode_with_gemini_only(self):
"""Test auto mode behavior when only Gemini API key is available"""

17
tests/test_openai_provider.py
View File

@@ -50,7 +50,7 @@ class TestOpenAIProvider:
assert provider.validate_model_name("o4-mini") is True
assert provider.validate_model_name("gpt-5") is True
assert provider.validate_model_name("gpt-5-mini") is True
assert provider.validate_model_name("gpt-5.1") is True
assert provider.validate_model_name("gpt-5.2") is True
assert provider.validate_model_name("gpt-5.1-codex") is True
assert provider.validate_model_name("gpt-5.1-codex-mini") is True
@@ -62,6 +62,7 @@ class TestOpenAIProvider:
assert provider.validate_model_name("gpt5") is True
assert provider.validate_model_name("gpt5-mini") is True
assert provider.validate_model_name("gpt5mini") is True
assert provider.validate_model_name("gpt5.2") is True
assert provider.validate_model_name("gpt5.1") is True
assert provider.validate_model_name("gpt5.1-codex") is True
assert provider.validate_model_name("codex-mini") is True
@@ -83,7 +84,8 @@ class TestOpenAIProvider:
assert provider._resolve_model_name("gpt5") == "gpt-5"
assert provider._resolve_model_name("gpt5-mini") == "gpt-5-mini"
assert provider._resolve_model_name("gpt5mini") == "gpt-5-mini"
assert provider._resolve_model_name("gpt5.1") == "gpt-5.1"
assert provider._resolve_model_name("gpt5.2") == "gpt-5.2"
assert provider._resolve_model_name("gpt5.1") == "gpt-5.2"
assert provider._resolve_model_name("gpt5.1-codex") == "gpt-5.1-codex"
assert provider._resolve_model_name("codex-mini") == "gpt-5.1-codex-mini"
@@ -95,7 +97,8 @@ class TestOpenAIProvider:
assert provider._resolve_model_name("o4-mini") == "o4-mini"
assert provider._resolve_model_name("gpt-5") == "gpt-5"
assert provider._resolve_model_name("gpt-5-mini") == "gpt-5-mini"
assert provider._resolve_model_name("gpt-5.1") == "gpt-5.1"
assert provider._resolve_model_name("gpt-5.2") == "gpt-5.2"
assert provider._resolve_model_name("gpt-5.1") == "gpt-5.2"
assert provider._resolve_model_name("gpt-5.1-codex") == "gpt-5.1-codex"
assert provider._resolve_model_name("gpt-5.1-codex-mini") == "gpt-5.1-codex-mini"
@@ -158,12 +161,12 @@ class TestOpenAIProvider:
assert capabilities.supports_function_calling is True
assert capabilities.supports_temperature is True
def test_get_capabilities_gpt51(self):
"""Test GPT-5.1 capabilities reflect new metadata."""
def test_get_capabilities_gpt52(self):
"""Test GPT-5.2 capabilities reflect new metadata."""
provider = OpenAIModelProvider("test-key")
capabilities = provider.get_capabilities("gpt-5.1")
assert capabilities.model_name == "gpt-5.1"
capabilities = provider.get_capabilities("gpt-5.2")
assert capabilities.model_name == "gpt-5.2"
assert capabilities.supports_streaming is True
assert capabilities.supports_function_calling is True
assert capabilities.supports_json_mode is True

10
tests/test_per_tool_model_defaults.py
View File

@@ -133,8 +133,8 @@ class TestModelSelection:
ModelProviderRegistry.register_provider(ProviderType.OPENAI, OpenAIModelProvider)
model = ModelProviderRegistry.get_preferred_fallback_model(ToolModelCategory.FAST_RESPONSE)
# OpenAI now prefers gpt-5.1 for fast response (based on our new preference order)
assert model == "gpt-5.1"
# OpenAI now prefers gpt-5.2 for fast response (based on our new preference order)
assert model == "gpt-5.2"
def test_fast_response_with_gemini_only(self):
"""Test FAST_RESPONSE prefers flash when only Gemini is available."""
@@ -167,8 +167,8 @@ class TestModelSelection:
ModelProviderRegistry.register_provider(ProviderType.OPENAI, OpenAIModelProvider)
model = ModelProviderRegistry.get_preferred_fallback_model(ToolModelCategory.BALANCED)
# OpenAI prefers gpt-5.1 for balanced (based on our new preference order)
assert model == "gpt-5.1"
# OpenAI prefers gpt-5.2 for balanced (based on our new preference order)
assert model == "gpt-5.2"
def test_no_category_uses_balanced_logic(self):
"""Test that no category specified uses balanced logic."""
@@ -209,7 +209,7 @@ class TestFlexibleModelSelection:
"env": {"OPENAI_API_KEY": "test-key"},
"provider_type": ProviderType.OPENAI,
"category": ToolModelCategory.FAST_RESPONSE,
"expected": "gpt-5.1", # Based on new preference order
"expected": "gpt-5.2", # Based on new preference order
},
]

8
tests/test_providers.py
View File

@@ -209,7 +209,7 @@ class TestOpenAIProvider:
assert provider.validate_model_name("o4-mini")
assert provider.validate_model_name("o4mini")
assert provider.validate_model_name("o4-mini")
assert provider.validate_model_name("gpt-5.1")
assert provider.validate_model_name("gpt-5.2")
assert provider.validate_model_name("gpt-5.1-codex")
assert provider.validate_model_name("gpt-5.1-codex-mini")
assert not provider.validate_model_name("gpt-4o")
@@ -223,11 +223,11 @@ class TestOpenAIProvider:
for alias in aliases:
assert not provider.get_capabilities(alias).supports_extended_thinking
def test_gpt51_family_capabilities(self):
"""Ensure GPT-5.1 family exposes correct capability flags."""
def test_gpt52_family_capabilities(self):
"""Ensure GPT-5.2 base model exposes correct capability flags."""
provider = OpenAIModelProvider(api_key="test-key")
base = provider.get_capabilities("gpt-5.1")
base = provider.get_capabilities("gpt-5.2")
assert base.supports_streaming
assert base.allow_code_generation

7
tests/test_supported_models_aliases.py
View File

@@ -54,7 +54,7 @@ class TestSupportedModelsAliases:
assert "o3mini" in provider.MODEL_CAPABILITIES["o3-mini"].aliases
assert "o3pro" in provider.MODEL_CAPABILITIES["o3-pro"].aliases
assert "gpt4.1" in provider.MODEL_CAPABILITIES["gpt-4.1"].aliases
assert "gpt5.1" in provider.MODEL_CAPABILITIES["gpt-5.1"].aliases
assert "gpt5.2" in provider.MODEL_CAPABILITIES["gpt-5.2"].aliases
assert "gpt5.1-codex" in provider.MODEL_CAPABILITIES["gpt-5.1-codex"].aliases
assert "codex-mini" in provider.MODEL_CAPABILITIES["gpt-5.1-codex-mini"].aliases
@@ -64,14 +64,15 @@ class TestSupportedModelsAliases:
assert provider._resolve_model_name("o3pro") == "o3-pro" # o3pro resolves to o3-pro
assert provider._resolve_model_name("o4mini") == "o4-mini"
assert provider._resolve_model_name("gpt4.1") == "gpt-4.1" # gpt4.1 resolves to gpt-4.1
assert provider._resolve_model_name("gpt5.1") == "gpt-5.1"
assert provider._resolve_model_name("gpt5.2") == "gpt-5.2"
assert provider._resolve_model_name("gpt5.1") == "gpt-5.2"
assert provider._resolve_model_name("gpt5.1-codex") == "gpt-5.1-codex"
assert provider._resolve_model_name("codex-mini") == "gpt-5.1-codex-mini"
# Test case insensitive resolution
assert provider._resolve_model_name("Mini") == "gpt-5-mini" # mini -> gpt-5-mini now
assert provider._resolve_model_name("O3MINI") == "o3-mini"
assert provider._resolve_model_name("Gpt5.1") == "gpt-5.1"
assert provider._resolve_model_name("Gpt5.1") == "gpt-5.2"
def test_xai_provider_aliases(self):
"""Test XAI provider's alias structure."""