Skip to content

Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.

fix(vlm): treat gpt-6 and later OpenAI models as reasoning models (#5397) (#5398)

Mohammad Hijjawia42752ee38f8
diff --git a/openviking/models/vlm/backends/openai_vlm.py b/openviking/models/vlm/backends/openai_vlm.py
index 5bcec67d3..7f4d6b1bd 100644
--- a/openviking/models/vlm/backends/openai_vlm.py
+++ b/openviking/models/vlm/backends/openai_vlm.py
@@ -4,6 +4,7 @@
import base64
import json
+import re
import time
from pathlib import Path
from typing import Any, Dict, List, Optional, Union
@@ -41,7 +42,8 @@ _DASHSCOPE_HOSTS = {
# These names select existing OpenAI request defaults, not general reasoning capability.
-_OPENAI_REASONING_MODEL_PREFIXES = ("gpt-5", "o1", "o3", "o4")
+# gpt-5 and later generations (gpt-6, gpt-7, ...) share the o-series request contract.
+_OPENAI_REASONING_MODEL_PATTERN = re.compile(r"(?:gpt-(?:[5-9]|[1-9]\d)|o[134])")
def _build_openai_client_kwargs(
@@ -181,7 +183,7 @@ class OpenAIVLM(VLMBase):
self, kwargs: Dict[str, Any], thinking: bool, max_tokens: Optional[int] = None
) -> None:
"""Apply model defaults, explicit settings, and provider-specific body fields."""
- if kwargs["model"].lower().startswith(_OPENAI_REASONING_MODEL_PREFIXES):
+ if _OPENAI_REASONING_MODEL_PATTERN.match(kwargs["model"].lower()):
max_tokens_param = "max_completion_tokens"
kwargs["reasoning_effort"] = "low"
else:
diff --git a/tests/unit/test_vlm_reasoning_models.py b/tests/unit/test_vlm_reasoning_models.py
index 976a6201c..2fc540423 100644
--- a/tests/unit/test_vlm_reasoning_models.py
+++ b/tests/unit/test_vlm_reasoning_models.py
@@ -2,6 +2,8 @@
# SPDX-License-Identifier: AGPL-3.0
"""Tests for OpenAI request defaults and explicitly configured reasoning effort."""
+import pytest
+
from openviking.models.vlm.backends.openai_vlm import OpenAIVLM
from openviking_cli.utils.config.vlm_config import VLMConfig
@@ -9,11 +11,12 @@ from openviking_cli.utils.config.vlm_config import VLMConfig
class TestOpenAITextCompletionParams:
"""OpenAI defaults must not gate compatible-provider configuration."""
- def test_gpt5_mini_uses_max_completion_tokens(self):
+ @pytest.mark.parametrize("model", ["gpt-5-mini", "gpt-6-luna", "GPT-7"])
+ def test_reasoning_families_use_max_completion_tokens(self, model):
vlm = OpenAIVLM(
{
"api_key": "sk-test",
- "model": "gpt-5-mini",
+ "model": model,
"api_base": "https://api.openai.com/v1",
"max_tokens": 512,
}