Skip to content

Build AI methods with your coding agent, then run them anywhere: as an MCP for chatbots, as a webapp for people, or via API for your software.

Update LLM deck defaults, fix object list e2e test, add bedrock token auth changelog

Louis Choquel8c1370a5014c
diff --git a/CHANGELOG.md b/CHANGELOG.md
index e3e55f7a4..d93d92514 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -9,6 +9,7 @@
- **`TEMPERATURE_UNSUPPORTED` constraint**: New listed constraint for models that reject sampling parameters entirely, checked in both Anthropic and OpenAI completions workers
- **claude-4.6-sonnet model**: Registered on anthropic, bedrock, and gateway backends
- **LLM deck cheap presets**: Added cheap variants for writing-factual, retrieval, and engineering-code presets, with retrieval tiers from `gemini-2.5-flash-lite` to `claude-4.7-opus`
+- **Bedrock bearer token authentication**: New `bedrock_access_variant` config option supports `"bedrock_token"` auth using `AWS_BEARER_TOKEN_BEDROCK` env var, alongside the existing `"aws_access"` method (default)
### Changed
diff --git a/pipelex/kit/configs/inference/deck/1_llm_deck.toml b/pipelex/kit/configs/inference/deck/1_llm_deck.toml
index e83bf0825..6e046e1b2 100644
--- a/pipelex/kit/configs/inference/deck/1_llm_deck.toml
+++ b/pipelex/kit/configs/inference/deck/1_llm_deck.toml
@@ -28,7 +28,7 @@ for_object = "@default-general"
[llm.aliases]
best-gpt = "gpt-5.2"
-best-claude = "claude-4.6-opus"
+best-claude = "claude-4.7-opus"
best-gemini = "gemini-3.0-pro"
best-mistral = "mistral-large"
@@ -41,8 +41,8 @@ default-large-context-code = "gemini-3.0-pro"
default-large-context-text = "gemini-2.5-flash"
default-small = "gpt-4o-mini"
default-small-structured = "gpt-4o-mini"
-default-small-vision = "gemini-2.5-flash-lite"
-default-small-creative = "gemini-2.5-flash-lite"
+default-small-vision = "gemini-3.0-flash-preview"
+default-small-creative = "gemini-3.0-flash-preview"
####################################################################################################
# LLM Presets
@@ -52,34 +52,36 @@ default-small-creative = "gemini-2.5-flash-lite"
# Writing
writing-factual = { model = "@default-premium", temperature = 0.1, description = "Factual writing with high accuracy" }
+writing-factual-cheap = { model = "@default-small", temperature = 0.1, description = "Cheap factual writing with high accuracy" }
writing-creative = { model = "@default-premium", temperature = 0.9, description = "Creative writing with high variability" }
+writing-creative-cheap = { model = "@default-small-creative", temperature = 0.9, description = "Cheap creative writing with high variability" }
# Retrieval
retrieval = { model = "@default-large-context-text", temperature = 0.1, description = "Data retrieval from large text corpora" }
+retrieval-cheap = { model = "gemini-2.5-flash-lite", temperature = 0.1, description = "Cheap data retrieval from large text corpora" }
+retrieval-premium = { model = "claude-4.7-opus", temperature = 0.1, description = "Premium data retrieval with highest accuracy" }
# Engineering
engineering-structured = { model = "@default-premium-structured", temperature = 0.2, description = "Structured engineering output (JSON, schemas)" }
engineering-code = { model = "@default-premium", temperature = 0.1, description = "Code generation and analysis" }
+engineering-code-cheap = { model = "claude-4.6-sonnet", temperature = 0.1, description = "Cheap code generation and analysis" }
+engineering-code-cheaper = { model = "claude-4.5-haiku", temperature = 0.1, description = "Cheapest code generation and analysis" }
engineering-codebase-analysis = { model = "@best-gemini", temperature = 0.1, description = "Large codebase analysis" }
# Vision
-vision = { model = "@default-premium-vision", temperature = 0.5, description = "Vision language model for understanding images" }
-vision-cheap = { model = "@default-small-vision", temperature = 0.5, description = "Budget vision model for simple image tasks" }
+vision = { model = "@default-premium-vision", temperature = 0.5, description = "Understanding and interpreting images" }
+vision-cheap = { model = "@default-small-vision", temperature = 0.5, description = "Cheap image understanding for simple tasks" }
vision-diagram = { model = "@default-premium-vision", temperature = 0.3, description = "Diagram and chart interpretation" }
vision-table = { model = "@default-premium-vision", temperature = 0.3, description = "Table extraction from images" }
# Image generation prompting
img-gen-prompting = { model = "@default-premium", temperature = 0.5, description = "Crafting image generation prompts" }
-img-gen-prompting-cheap = { model = "@default-small", temperature = 0.5, description = "Budget image prompt generation" }
+img-gen-prompting-cheap = { model = "@default-small-creative", temperature = 0.5, description = "Cheap image prompt generation" }
# Reasoning
deep-analysis = { model = "@default-premium", temperature = 0.1, reasoning_effort = "high", description = "Deep reasoning and analysis" }
quick-reasoning = { model = "@default-premium", temperature = 0.3, reasoning_effort = "low", description = "Quick reasoning for simple tasks" }
-# Builder (isolated presets for the pipeline builder)
-pipe-builder-engineering = { model = "claude-4.6-opus", temperature = 0.2, description = "Builder: structured engineering output" }
-pipe-builder-img-gen-prompting = { model = "claude-4.5-sonnet", temperature = 0.7, description = "Builder: crafting image generation prompts" }
-
# Testing
testing-text = { model = "@default-small", temperature = 0.5, description = "Testing preset for text generation" }
testing-structured = { model = "@default-small-structured", temperature = 0.1, description = "Testing preset for structured output" }
diff --git a/pipelex/pipelex.toml b/pipelex/pipelex.toml
index 4150be5df..454307fa5 100644
--- a/pipelex/pipelex.toml
+++ b/pipelex/pipelex.toml
@@ -117,7 +117,7 @@ keywords_to_hilight = []
[pipelex.aws_config]
api_key_method = "env"
-bedrock_access_variant = "bedrock_token"
+bedrock_access_variant = "aws_access"
####################################################################################################
# Cogt inference config
diff --git a/tests/e2e/pipelex/pipes/pipe_operators/pipe_llm/pipe_llm_object_list.mthds b/tests/e2e/pipelex/pipes/pipe_operators/pipe_llm/pipe_llm_object_list.mthds
index 8e434ec55..8eebc90ca 100644
--- a/tests/e2e/pipelex/pipes/pipe_operators/pipe_llm/pipe_llm_object_list.mthds
+++ b/tests/e2e/pipelex/pipes/pipe_operators/pipe_llm/pipe_llm_object_list.mthds
@@ -14,13 +14,12 @@ inputs = { inspiration = "Text" }
output = "MoodboardPrompt[]"
model = "@default-small-creative"
prompt = """
-You are a fashion art director creating a moodboard. Given the following fashion inspiration, generate exactly 5 distinct image generation prompts. Each prompt should capture a different visual facet of the aesthetic:
+You are a fashion art director creating a moodboard. Given the following fashion inspiration, generate exactly 3 distinct image generation prompts.
+Each prompt should capture a different visual facet of the aesthetic:
1. A hero outfit or look
2. A texture or fabric close-up
3. A color palette scene
-4. A setting or environment that evokes the mood
-5. An accessory or detail shot
Fashion inspiration:
@inspiration
diff --git a/tests/e2e/pipelex/pipes/pipe_operators/pipe_llm/test_pipe_llm_object_list.py b/tests/e2e/pipelex/pipes/pipe_operators/pipe_llm/test_pipe_llm_object_list.py
index 6e43a8f6f..1f36ca0f1 100644
--- a/tests/e2e/pipelex/pipes/pipe_operators/pipe_llm/test_pipe_llm_object_list.py
+++ b/tests/e2e/pipelex/pipes/pipe_operators/pipe_llm/test_pipe_llm_object_list.py
@@ -11,7 +11,7 @@ from pipelex.pipeline.runner import PipelexRunner
@pytest.mark.llm
@pytest.mark.inference
@pytest.mark.dry_runnable
[email protected](loop_scope="class")
class TestPipeLLMObjectList:
async def test_craft_prompts(self, pipe_run_mode: PipeRunMode) -> None:
"""Test a PipeLLM pipe that generates a list of structured MoodboardPrompt objects."""
@@ -29,6 +29,6 @@ class TestPipeLLMObjectList:
assert pipeline_response.pipe_output.main_stuff is not None
items = pipeline_response.pipe_output.main_stuff_as_list(item_type=TextContent)
- assert len(items) == 5
+ assert len(items) == 3
pretty_print(items, title="Moodboard prompts")