feat(together-ai/LiquidAI/LFM2.5-8B-A1B): add new models [bot]#1733
feat(together-ai/LiquidAI/LFM2.5-8B-A1B): add new models [bot]#1733models-bot[bot] wants to merge 2 commits into
Conversation
|
/test-models |
Gateway test results
Failures (2)
ErrorCode snippetfrom openai import OpenAI
client = OpenAI(api_key="***", base_url="https://internal.devtest.truefoundry.tech/api/llm")
response = client.chat.completions.create(
model="test-v2-together-ai/LiquidAI-LFM2.5-8B-A1B",
messages=[
{"role": "user", "content": "How to calculate 3^3^3^3? Think step by step and show all reasoning."},
],
reasoning_effort="medium",
stream=False,
)
_usage = getattr(response, "usage", None)
_reasoning_detected = False
_choices = getattr(response, "choices", None)
if _choices and len(_choices) > 0:
_message = getattr(_choices[0], "message", None)
else:
_message = None
if _message and getattr(_message, "content", None) is not None:
print(_message.content)
if _usage is not None:
_output_token_details = getattr(_usage, "completion_tokens_details", None)
if _output_token_details and getattr(_output_token_details, "reasoning_tokens", 0) > 0:
_reasoning_detected = True
elif getattr(_usage, "reasoning", None) is not None:
_reasoning_detected = True
if getattr(_message, "reasoning_content", None) is not None:
_reasoning_detected = True
elif getattr(_message, "reasoning", None) is not None:
_reasoning_detected = True
if not _reasoning_detected:
print("Response: ", response)
raise Exception("VALIDATION FAILED: reasoning - no reasoning information in response")
print("VALIDATION: reasoning SUCCESS")Output
ErrorCode snippetfrom openai import OpenAI
client = OpenAI(api_key="***", base_url="https://internal.devtest.truefoundry.tech/api/llm")
response = client.chat.completions.create(
model="test-v2-together-ai/LiquidAI-LFM2.5-8B-A1B",
messages=[
{"role": "user", "content": "How to calculate 3^3^3^3? Think step by step and show all reasoning."},
],
reasoning_effort="medium",
stream=True,
)
_reasoning_detected = False
for chunk in response:
if chunk.choices and len(chunk.choices) > 0:
delta = chunk.choices[0].delta
if delta.content is not None:
print(delta.content, end="", flush=True)
if getattr(delta, "reasoning_content", None) is not None:
_reasoning_detected = True
if getattr(delta, "reasoning", None) is not None:
_reasoning_detected = True
_usage = getattr(chunk, "usage", None)
if _usage is not None:
_details = getattr(_usage, "completion_tokens_details", None)
if _details and getattr(_details, "reasoning_tokens", 0) > 0:
_reasoning_detected = True
if not _reasoning_detected:
raise Exception("VALIDATION FAILED: reasoning stream - no reasoning information in stream")
print("\nVALIDATION: reasoning stream SUCCESS")OutputSuccesses (2)
Output
Output |
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes using default effort and found 3 potential issues.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit 473891d. Configure here.
| output_cost_per_token: 1.2e-7 | ||
| region: "*" | ||
| limits: | ||
| context_window: 32768 |
There was a problem hiding this comment.
Wrong context window limit
Medium Severity
limits.context_window is set to 32768, but the model’s cited Liquid AI release describes a 128K context window. Downstream consumers of this registry may cap or validate prompts at 32K tokens and block or truncate valid long-context use.
Reviewed by Cursor Bugbot for commit 473891d. Configure here.
| - https://www.liquid.ai/blog/lfm2-5-8b-a1b | ||
| status: active | ||
| supportedModes: | ||
| - chat |
There was a problem hiding this comment.
Missing tool calling features
Medium Severity
The config has no features block, so capabilities like function_calling and tool_choice are not declared. LFM2.5-8B-A1B is positioned for tool calling, and the provider’s other Liquid entry lists those features explicitly.
Reviewed by Cursor Bugbot for commit 473891d. Configure here.
| status: active | ||
| supportedModes: | ||
| - chat | ||
| thinking: true |
There was a problem hiding this comment.
Incorrect thinking flag set
Low Severity
thinking: true marks extended reasoning support, but LFM2.5 is documented for tool calling rather than step-by-step reasoning, and the other Liquid model on Together omits this flag.
Reviewed by Cursor Bugbot for commit 473891d. Configure here.


Auto-generated by model-addition-agent for
together-ai/LiquidAI/LFM2.5-8B-A1B.Note
Low Risk
Single new YAML model entry with no changes to application logic or security-sensitive paths.
Overview
Adds a new Together AI provider definition for
LiquidAI/LFM2.5-8B-A1B, registering it as an active, serverless chat model with 32k context, text in/out, per-token pricing, andthinking: true.This is catalog/metadata only (no runtime or routing logic changes).
Reviewed by Cursor Bugbot for commit 473891d. Bugbot is set up for automated code reviews on this repo. Configure here.