unique_toolkit 2026.44.0.dev2__py3-none-any.whl → 2026.44.0.dev3__py3-none-any.whl

unique_toolkit/language_model/infos.py

@@ -176,6 +176,7 @@ class LanguageModelName(StrEnum):
     VERTEX_CLAUDE_OPUS_5_5 = "litellm:vertex-claude-opus-5-5"
     VERTEX_CLAUDE_FABLE_5 = "litellm:vertex-claude-fable-5"
     VERTEX_GEMINI_3_8_FLASH = "litellm:vertex-gemini-3-8-flash"
+    VERTEX_GEMINI_4_ARGON = "litellm:vertex-gemini-4-argon"
 
 
 class EncoderName(StrEnum):
@@ -2487,11 +2488,10 @@ class LanguageModelInfo(BaseModel):
                     default_options={"reasoning_effort": "low"},
                     supported_reasoning_efforts=["low", "high"],
                 )
-            # Day-0 placeholder so the model can be selected before general
-            # availability. Limits, cutoff, thinking levels, and the API model
-            # id are provisional. Update this entry from the official model
-            # card once Argon is published.
-            case LanguageModelName.GEMINI_4_ARGON:
+            case (
+                LanguageModelName.GEMINI_4_ARGON
+                | LanguageModelName.VERTEX_GEMINI_4_ARGON
+            ):
                 return cls(
                     name=model_name,
                     capabilities=[
@@ -2506,21 +2506,16 @@ class LanguageModelInfo(BaseModel):
                     version="gemini-4-argon",
                     encoder_name=EncoderName.O200K_BASE,  # TODO: Update encoder with litellm
                     token_limits=LanguageModelTokenLimits(
-                        # GraphWalks evaluates contexts from 256k through 1M tokens.
-                        # https://deepmind.google/models/gemini/
+                        # Unified 1,048,576 input/output context window; 65,536 single-request output cap.
+                        # https://ai.google.dev/gemini-api/docs/models/gemini-4-argon
+                        # https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/gemini/4-argon
                         token_limit_input=1_048_576,
                         token_limit_output=65_536,
                     ),
-                    # Google has not published a knowledge cutoff for Argon.
-                    info_cutoff_at=None,
-                    # https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/
-                    published_at=date(2026, 9, 30),
-                    # thinkingLevel is recommended for Gemini 3 and later. Argon is
-                    # not yet on the public thinking-level table; evals use the
-                    # highest thinking setting. Mirror Gemini 3.8 Flash (low,
-                    # medium, high; medium default) until Argon docs land.
-                    # https://ai.google.dev/gemini-api/docs/thinking
-                    # https://deepmind.google/models/evals-methodology/gemini-4-argon
+                    info_cutoff_at=date(2026, 3, day=1),
+                    published_at=date(2026, 10, 13),
+                    # Supports thinking_level low, medium (default), and high.
+                    # https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/guides/gemini-4-argon
                     default_options={"reasoning_effort": "medium"},
                     supported_reasoning_efforts=["low", "medium", "high"],
                 )

{unique_toolkit-2026.44.0.dev2.dist-info → unique_toolkit-2026.44.0.dev3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: unique-toolkit
-Version: 2026.44.0.dev2
+Version: 2026.44.0.dev3
 Summary: 
 Author: Cedric Klinkert, Martin Fadler, Sadique Sheik, Fabian Schläpfer, Pascal Hauri, Ejiroghene Oki-Peter, Rami Azouz, Ricardo Visini, Sanzio Monti
 Author-email: Cedric Klinkert <cedric.klinkert@unique.ch>, Martin Fadler <martin.fadler@unique.ch>, Sadique Sheik <sadique@unique.ch>, Fabian Schläpfer <fabian@unique.ch>, Pascal Hauri <pascal@unique.ch>, Ejiroghene Oki-Peter <oki-peter.ejiroghene@unique.ai>, Rami Azouz <rami.ext@unique.ch>, Ricardo Visini <ricardo.visini@unique.ch>, Sanzio Monti <sanzio.monti@unique.ai>

{unique_toolkit-2026.44.0.dev2.dist-info → unique_toolkit-2026.44.0.dev3.dist-info}/RECORD

@@ -397,7 +397,7 @@ unique_toolkit/language_model/builder.py,sha256=KwBzsD6-fvQnB4p2I7M-5Tk6j4tmQ973
 unique_toolkit/language_model/constants.py,sha256=B-topqW0r83dkC_25DeQfnPk3n53qzIHUCBS7YJ0-1U,119
 unique_toolkit/language_model/default_language_model.py,sha256=mcyuFs_V6tp52C8xJSHiEHO4kFGj6xed1gvozvCRD7k,1181
 unique_toolkit/language_model/functions.py,sha256=W6JC4LUh-LZ9waDgph--xVmKeQ7Ls6NRAqDx9DwQdeY,26691
-unique_toolkit/language_model/infos.py,sha256=nDntUjD5NhByYthSKLDhoJ54ThnwzDbeclxmInN00ic,182747
+unique_toolkit/language_model/infos.py,sha256=f1xQQ8BvLG9jPsEt2vdSYzxVvaKIkem-Y6T4_n4tT_c,182327
 unique_toolkit/language_model/invocation_stats.py,sha256=CQbpUz394IsVd8eNTB1gPSMt5f0UXmkN1amNsqtAW6U,5857
 unique_toolkit/language_model/model_costs.py,sha256=ShuDAzMew1fUkJ3tzPgU5Zqjp6ivUZaG97BFh93tEn4,6358
 unique_toolkit/language_model/prompt.py,sha256=JSawaLjQg3VR-E2fK8engFyJnNdk21zaO8pPIodzN4Q,3991
@@ -426,6 +426,6 @@ unique_toolkit/short_term_memory/service.py,sha256=2JMBEBn6GoyK-a0xvdcWpPft-3C74
 unique_toolkit/smart_rules/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 unique_toolkit/smart_rules/compile.py,sha256=vRe0aMJWuVIK-VsqomgxQImzke7MDmwgSxxJfYCa7zw,1386
 unique_toolkit/test_utilities/events.py,sha256=FLrhMh0A_mQutAUjoD0FdMcVQ4Gq31r35OVGzRjvLro,6760
-unique_toolkit-2026.44.0.dev2.dist-info/WHEEL,sha256=eh7sammvW2TypMMMGKgsM83HyA_3qQ5Lgg3ynoecH3M,79
-unique_toolkit-2026.44.0.dev2.dist-info/METADATA,sha256=6M-EKmFU3NXNxhKdv9lM_8bxtQKHfgZfdddbk74v4AM,7821
-unique_toolkit-2026.44.0.dev2.dist-info/RECORD,,
+unique_toolkit-2026.44.0.dev3.dist-info/WHEEL,sha256=eh7sammvW2TypMMMGKgsM83HyA_3qQ5Lgg3ynoecH3M,79
+unique_toolkit-2026.44.0.dev3.dist-info/METADATA,sha256=WmTxbFMxjCN7yCsDn17P-sDhnxSDxcQL-sSn6LgvBGg,7821
+unique_toolkit-2026.44.0.dev3.dist-info/RECORD,,