lm-deluge 0.0.88__py3-none-any.whl → 0.0.89__py3-none-any.whl

lm_deluge/api_requests/gemini.py

@@ -52,6 +52,7 @@ async def _build_gemini_request(
 
     # Handle reasoning models (thinking)
     is_gemini_3 = "gemini-3" in model.name.lower()
+    is_gemini_3_flash = "gemini-3-flash" in model.name.lower()
     if is_gemini_3:
         # gemini3 MUST think
         if not sampling_params.reasoning_effort:
@@ -62,13 +63,24 @@ async def _build_gemini_request(
             if effort_key == "xhigh":
                 maybe_warn("WARN_XHIGH_TO_HIGH", model_name=model.name)
                 effort_key = "high"
-            level_map = {
-                "none": "low",
-                "minimal": "low",
-                "low": "low",
-                "medium": "high",  # change when supported
-                "high": "high",
-            }
+            if is_gemini_3_flash:
+                # Flash supports minimal, low, medium, high
+                level_map = {
+                    "none": "low",
+                    "minimal": "minimal",
+                    "low": "low",
+                    "medium": "medium",
+                    "high": "high",
+                }
+            else:
+                # Pro only supports low, high
+                level_map = {
+                    "none": "low",
+                    "minimal": "low",
+                    "low": "low",
+                    "medium": "high",
+                    "high": "high",
+                }
             effort = level_map[effort_key]
         thinking_config = {"thinkingLevel": effort}
         request_json["generationConfig"]["thinkingConfig"] = thinking_config

lm_deluge/models/google.py

@@ -153,6 +153,19 @@ GOOGLE_MODELS = {
         # Note: >200k tokens pricing is $4/$18 per million
         "reasoning_model": True,
     },
+    "gemini-3-flash-preview": {
+        "id": "gemini-3-flash-preview",
+        "name": "gemini-3-flash-preview",
+        "api_base": "https://generativelanguage.googleapis.com/v1alpha",
+        "api_key_env_var": "GEMINI_API_KEY",
+        "supports_json": True,
+        "supports_logprobs": False,
+        "api_spec": "gemini",
+        "input_cost": 0.5,
+        "cached_input_cost": 0.125,  # estimated
+        "output_cost": 3.0,
+        "reasoning_model": True,
+    },
     # Gemini 2.5 Computer Use model
     "gemini-2.5-computer-use": {
         "id": "gemini-2.5-computer-use",

{lm_deluge-0.0.88.dist-info → lm_deluge-0.0.89.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lm_deluge
-Version: 0.0.88
+Version: 0.0.89
 Summary: Python utility for using LLM API models.
 Author-email: Benjamin Anderson <ben@trytaylor.ai>
 Requires-Python: >=3.10

{lm_deluge-0.0.88.dist-info → lm_deluge-0.0.89.dist-info}/RECORD

@@ -21,7 +21,7 @@ lm_deluge/api_requests/base.py,sha256=05j5nrZhgNon2YRFXT_L-yVXKlvdodwOJan6Z6WpSp
 lm_deluge/api_requests/bedrock.py,sha256=mY1xTvgfCLyqLlfFFmu_baKgkVq1Df1_MJXeN_G1jWQ,15597
 lm_deluge/api_requests/chat_reasoning.py,sha256=sJvstvKFqsSBUjYcwxzGt2_FH4cEp3Z6gKcBPyPjGwk,236
 lm_deluge/api_requests/common.py,sha256=BZ3vRO5TB669_UsNKugkkuFSzoLHOYJIKt4nV4sf4vc,422
-lm_deluge/api_requests/gemini.py,sha256=_L1V1O7N70rvnp28kXfCaWXluQXEBfMbvSk3S7SrZcI,12730
+lm_deluge/api_requests/gemini.py,sha256=osVbTVCdVef6sfE0XsG13olsnDHB2J70_c6G2IP-sIg,13189
 lm_deluge/api_requests/mistral.py,sha256=8JZP2CDf1XZfaPcTk0WS4q-VfYYj58ptpoH8LD3MQG4,4528
 lm_deluge/api_requests/openai.py,sha256=cEoMpMPKrVTz2Zjm5pdD8sQpOHDB9O8ndwf9TGqLPcA,29889
 lm_deluge/api_requests/response.py,sha256=vG194gAH5p7ulpNy4qy5Pryfb1p3ZV21-YGoj__ru3E,7436
@@ -38,7 +38,7 @@ lm_deluge/models/cerebras.py,sha256=u2FMXJF6xMr0euDRKLKMo_NVTOcvSrrEpehbHr8sSeE,
 lm_deluge/models/cohere.py,sha256=iXjYtM6jy_YL73Op8OfNsrMNopwae9y-Sw-4vF9cEBw,3406
 lm_deluge/models/deepseek.py,sha256=b5t_ep6fE-2cKD2mmImBaLcJUbYrfizYnjG96sfKNTk,2072
 lm_deluge/models/fireworks.py,sha256=yvt2Ggzye4aUqCqY74ta67Vu7FrQaLFjdFtN4P7D-dc,638
-lm_deluge/models/google.py,sha256=IDePlNOvF0lvpv3UhkUD8g30TUJqoaJHQGzTglyGg80,6560
+lm_deluge/models/google.py,sha256=ka5tBdpYT26zxsE1wzrl81Rt5XgpBQkPIOXOq5pSpRE,7033
 lm_deluge/models/grok.py,sha256=rSvN3fKiO_WPNa5O_TzVTDj9-RqqjeXFBiC9OAcGZ4Q,3340
 lm_deluge/models/groq.py,sha256=Mi5WE1xOBGoZlymD0UN6kzhH_NOmfJYU4N2l-TO0Z8Q,2552
 lm_deluge/models/kimi.py,sha256=B_ZL4_0q6hS1VVskBWlBR569nNSjC8RgA2lj1eCjRRE,1183
@@ -110,8 +110,8 @@ lm_deluge/util/schema.py,sha256=q6uwhA4s1lM2dHT1Kwc46E7OY1VecMOtTEI0PTFn6tA,1320
 lm_deluge/util/spatial.py,sha256=BsF_UKhE-x0xBirc-bV1xSKZRTUhsOBdGqsMKme20C8,4099
 lm_deluge/util/validation.py,sha256=hz5dDb3ebvZrZhnaWxOxbNSVMI6nmaOODBkk0htAUhs,1575
 lm_deluge/util/xml.py,sha256=Ft4zajoYBJR3HHCt2oHwGfymGLdvp_gegVmJ-Wqk4Ck,10547
-lm_deluge-0.0.88.dist-info/licenses/LICENSE,sha256=uNNXGXPCw2TC7CUs7SEBkA-Mz6QBQFWUUEWDMgEs1dU,1058
-lm_deluge-0.0.88.dist-info/METADATA,sha256=SNJZGw7OHpfMUVG5CvYyeBQ0lNiLUg1FbZis2XXj2eg,14057
-lm_deluge-0.0.88.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-lm_deluge-0.0.88.dist-info/top_level.txt,sha256=hqU-TJX93yBwpgkDtYcXyLr3t7TLSCCZ_reytJjwBaE,10
-lm_deluge-0.0.88.dist-info/RECORD,,
+lm_deluge-0.0.89.dist-info/licenses/LICENSE,sha256=uNNXGXPCw2TC7CUs7SEBkA-Mz6QBQFWUUEWDMgEs1dU,1058
+lm_deluge-0.0.89.dist-info/METADATA,sha256=JHgKoBFnOQJ8y1HTpeWhFYBi4H4SJoJ3ZwsfGMJFc_U,14057
+lm_deluge-0.0.89.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+lm_deluge-0.0.89.dist-info/top_level.txt,sha256=hqU-TJX93yBwpgkDtYcXyLr3t7TLSCCZ_reytJjwBaE,10
+lm_deluge-0.0.89.dist-info/RECORD,,