Spaces:
Sleeping
Sleeping
Commit Β·
68a6539
1
Parent(s): d4090f3
feat: upgrade to GPT-5 series, increase default max_tokens to 8192
Browse files- OpenAI models: gpt-5 and gpt-5-mini only (removed GPT-4 series)
- Default max_tokens raised from 2048 to 8192 across all providers
- Added o4-mini to reasoning model detection
- Updated setup wizard fallback lists and defaults
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- setup.py +2 -2
- src/llm/__init__.py +2 -2
- src/llm/anthropic.py +1 -1
- src/llm/gemini.py +1 -1
- src/llm/openai.py +4 -4
- src/verifier.py +1 -1
setup.py
CHANGED
|
@@ -34,7 +34,7 @@ def generate_config(
|
|
| 34 |
"provider": provider,
|
| 35 |
"model": model,
|
| 36 |
"temperature": 0.0,
|
| 37 |
-
"max_tokens":
|
| 38 |
},
|
| 39 |
"api_keys": {
|
| 40 |
"openai": "",
|
|
@@ -153,7 +153,7 @@ def run_wizard():
|
|
| 153 |
except Exception:
|
| 154 |
print(f" Warning: Could not validate API key for {provider}. Using default model list.")
|
| 155 |
fallback = {
|
| 156 |
-
"openai": ["gpt-
|
| 157 |
"anthropic": ["claude-sonnet-4-6", "claude-opus-4-6", "claude-haiku-4-5"],
|
| 158 |
"gemini": ["gemini-2.5-pro", "gemini-2.5-flash", "gemini-2.0-flash"],
|
| 159 |
}
|
|
|
|
| 34 |
"provider": provider,
|
| 35 |
"model": model,
|
| 36 |
"temperature": 0.0,
|
| 37 |
+
"max_tokens": 8192,
|
| 38 |
},
|
| 39 |
"api_keys": {
|
| 40 |
"openai": "",
|
|
|
|
| 153 |
except Exception:
|
| 154 |
print(f" Warning: Could not validate API key for {provider}. Using default model list.")
|
| 155 |
fallback = {
|
| 156 |
+
"openai": ["gpt-5", "gpt-5-mini"],
|
| 157 |
"anthropic": ["claude-sonnet-4-6", "claude-opus-4-6", "claude-haiku-4-5"],
|
| 158 |
"gemini": ["gemini-2.5-pro", "gemini-2.5-flash", "gemini-2.0-flash"],
|
| 159 |
}
|
src/llm/__init__.py
CHANGED
|
@@ -22,7 +22,7 @@ def generate(system_prompt: str, user_message: str, cfg: dict,
|
|
| 22 |
model = llm_cfg.get("model") or None
|
| 23 |
temperature = llm_cfg.get("temperature", 0.0)
|
| 24 |
if max_tokens is None:
|
| 25 |
-
max_tokens = llm_cfg.get("max_tokens",
|
| 26 |
# Validate parameters
|
| 27 |
try:
|
| 28 |
temperature = max(0.0, min(float(temperature), 2.0))
|
|
@@ -34,7 +34,7 @@ def generate(system_prompt: str, user_message: str, cfg: dict,
|
|
| 34 |
try:
|
| 35 |
max_tokens = max(1, min(int(max_tokens), 128000))
|
| 36 |
except (ValueError, TypeError):
|
| 37 |
-
max_tokens =
|
| 38 |
return PROVIDERS[provider].generate(
|
| 39 |
system_prompt=system_prompt, user_message=user_message,
|
| 40 |
api_key=api_key, model=model, temperature=temperature, max_tokens=max_tokens,
|
|
|
|
| 22 |
model = llm_cfg.get("model") or None
|
| 23 |
temperature = llm_cfg.get("temperature", 0.0)
|
| 24 |
if max_tokens is None:
|
| 25 |
+
max_tokens = llm_cfg.get("max_tokens", 8192)
|
| 26 |
# Validate parameters
|
| 27 |
try:
|
| 28 |
temperature = max(0.0, min(float(temperature), 2.0))
|
|
|
|
| 34 |
try:
|
| 35 |
max_tokens = max(1, min(int(max_tokens), 128000))
|
| 36 |
except (ValueError, TypeError):
|
| 37 |
+
max_tokens = 8192
|
| 38 |
return PROVIDERS[provider].generate(
|
| 39 |
system_prompt=system_prompt, user_message=user_message,
|
| 40 |
api_key=api_key, model=model, temperature=temperature, max_tokens=max_tokens,
|
src/llm/anthropic.py
CHANGED
|
@@ -3,7 +3,7 @@ SUPPORTED_MODELS = ["claude-sonnet-4-6", "claude-opus-4-6", "claude-haiku-4-5"]
|
|
| 3 |
|
| 4 |
def generate(system_prompt: str, user_message: str, api_key: str,
|
| 5 |
model: str = None, temperature: float = 0.0,
|
| 6 |
-
max_tokens: int =
|
| 7 |
model = model or "claude-sonnet-4-6"
|
| 8 |
from anthropic import Anthropic
|
| 9 |
client = Anthropic(api_key=api_key)
|
|
|
|
| 3 |
|
| 4 |
def generate(system_prompt: str, user_message: str, api_key: str,
|
| 5 |
model: str = None, temperature: float = 0.0,
|
| 6 |
+
max_tokens: int = 8192) -> str:
|
| 7 |
model = model or "claude-sonnet-4-6"
|
| 8 |
from anthropic import Anthropic
|
| 9 |
client = Anthropic(api_key=api_key)
|
src/llm/gemini.py
CHANGED
|
@@ -9,7 +9,7 @@ SUPPORTED_MODELS = [
|
|
| 9 |
|
| 10 |
def generate(system_prompt: str, user_message: str, api_key: str,
|
| 11 |
model: str = None, temperature: float = 0.0,
|
| 12 |
-
max_tokens: int =
|
| 13 |
from google import genai
|
| 14 |
|
| 15 |
model = model or SUPPORTED_MODELS[0]
|
|
|
|
| 9 |
|
| 10 |
def generate(system_prompt: str, user_message: str, api_key: str,
|
| 11 |
model: str = None, temperature: float = 0.0,
|
| 12 |
+
max_tokens: int = 8192) -> str:
|
| 13 |
from google import genai
|
| 14 |
|
| 15 |
model = model or SUPPORTED_MODELS[0]
|
src/llm/openai.py
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
"""OpenAI LLM provider."""
|
| 2 |
-
SUPPORTED_MODELS = ["gpt-
|
| 3 |
|
| 4 |
def generate(system_prompt: str, user_message: str, api_key: str,
|
| 5 |
model: str = None, temperature: float = 0.0,
|
| 6 |
-
max_tokens: int =
|
| 7 |
-
model = model or "gpt-
|
| 8 |
from openai import OpenAI
|
| 9 |
client = OpenAI(api_key=api_key)
|
| 10 |
-
is_reasoning = model in ("o1", "o3", "o4") or model.startswith(("o1-", "o3-", "o4-"))
|
| 11 |
try:
|
| 12 |
if is_reasoning:
|
| 13 |
# Reasoning models include thinking tokens in max_completion_tokens,
|
|
|
|
| 1 |
"""OpenAI LLM provider."""
|
| 2 |
+
SUPPORTED_MODELS = ["gpt-5", "gpt-5-mini"]
|
| 3 |
|
| 4 |
def generate(system_prompt: str, user_message: str, api_key: str,
|
| 5 |
model: str = None, temperature: float = 0.0,
|
| 6 |
+
max_tokens: int = 8192) -> str:
|
| 7 |
+
model = model or "gpt-5"
|
| 8 |
from openai import OpenAI
|
| 9 |
client = OpenAI(api_key=api_key)
|
| 10 |
+
is_reasoning = model in ("o1", "o3", "o4", "o4-mini") or model.startswith(("o1-", "o3-", "o4-"))
|
| 11 |
try:
|
| 12 |
if is_reasoning:
|
| 13 |
# Reasoning models include thinking tokens in max_completion_tokens,
|
src/verifier.py
CHANGED
|
@@ -336,7 +336,7 @@ def verify_and_respond(
|
|
| 336 |
bot_name = cfg.get("chatbot", {}).get("name", "ResearchBot")
|
| 337 |
domain = cfg.get("chatbot", {}).get("domain", "research")
|
| 338 |
llm_cfg = cfg.get("llm", {})
|
| 339 |
-
default_max = llm_cfg.get("max_tokens",
|
| 340 |
|
| 341 |
# ββ Load KB overview for general awareness (brief version) ββββββββββ
|
| 342 |
from src.kb_meta import load_kb_meta_brief
|
|
|
|
| 336 |
bot_name = cfg.get("chatbot", {}).get("name", "ResearchBot")
|
| 337 |
domain = cfg.get("chatbot", {}).get("domain", "research")
|
| 338 |
llm_cfg = cfg.get("llm", {})
|
| 339 |
+
default_max = llm_cfg.get("max_tokens", 8192)
|
| 340 |
|
| 341 |
# ββ Load KB overview for general awareness (brief version) ββββββββββ
|
| 342 |
from src.kb_meta import load_kb_meta_brief
|