python-fastllm 0.0.16__tar.gz → 0.0.18__tar.gz

{python_fastllm-0.0.16 → python_fastllm-0.0.18}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: python-fastllm
-Version: 0.0.16
+Version: 0.0.18
 Author-email: Kerem Turgutlu <keremturgutlu@gmail.com>
 License: Apache-2.0
 Project-URL: Repository, https://github.com/AnswerDotAI/fastllm

python_fastllm-0.0.18/fastllm/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.0.18"

{python_fastllm-0.0.16 → python_fastllm-0.0.18}/fastllm/acomplete.py

@@ -43,7 +43,8 @@ vendor_mapping = {
     "openrouter":   ('openai_chat', "https://openrouter.ai/api/v1", "OPENROUTER_API_KEY"),
     "together":     ('openai_chat', "https://api.together.xyz/v1", "TOGETHER_API_KEY"),
     "fireworks_ai": ('openai_chat', "https://api.fireworks.ai/inference/v1", "FIREWORKS_API_KEY"),
-    "qwen":         ('openai_chat', "https://dashscope.aliyuncs.com/compatible-mode/v1", "QWEN_API_KEY")
+    "qwen":         ('openai_chat', "https://dashscope.aliyuncs.com/compatible-mode/v1", "QWEN_API_KEY"),
+    "minimax":      ('anthropic', "https://api.minimax.io/anthropic", "MINIMAX_API_KEY")
 }
 
 # %% ../nbs/06_acomplete.ipynb #77d27ea7

{python_fastllm-0.0.16 → python_fastllm-0.0.18}/fastllm/chat.py

@@ -620,10 +620,11 @@ def _handle_stop_reason(res):
 # %% ../nbs/07_chat.ipynb #daf876f4
 class StopReasonCallback(ChatCallback):
     order = 40
-    async def after_acomplete(self):        
-        action, msg = _handle_stop_reason(self.turn_res)
-        if action == 'warning': add_warning(self.chat.turn_res, msg)
-        if False: yield
+    async def after_acomplete(self):
+        action,msg = _handle_stop_reason(self.turn_res)
+        if action != 'warning': return
+        add_warning(self.chat.turn_res, msg)
+        if self.stream: yield dict(text=f"warning: {msg}\n\n")
 
 # %% ../nbs/07_chat.ipynb #aa7630b2
 def _active_fence_langs(tool_schemas):

{python_fastllm-0.0.16 → python_fastllm-0.0.18}/fastllm/types.py

@@ -340,10 +340,12 @@ for model in ('kimi-k2.5', 'kimi-k2.6'):
 register_model_info('gemini-3.1-flash-lite', vendor_name='gemini', base='gemini-3.1-flash-lite-preview')
 register_model_info('models/gemini-3.1-flash-lite', vendor_name='gemini', base='gemini-3.1-flash-lite-preview')
 
-for model in ('accounts/fireworks/models/kimi-k2p5', 'accounts/fireworks/models/kimi-k2p6'):
-    register_model_info(model, vendor_name='fireworks_ai', base=model.replace('k2p6', 'k2p5'),
-        supports_reasoning=True, supports_vision=True,
-        input_cost_per_token=0.95e-6, cache_read_input_token_cost=0.16e-6, output_cost_per_token=4.0e-6)
+register_model_info('accounts/fireworks/models/kimi-k2p5', vendor_name='fireworks_ai', base='accounts/fireworks/models/kimi-k2p5',
+    supports_reasoning=True, supports_vision=True,
+    input_cost_per_token=0.60e-6, cache_read_input_token_cost=0.10e-6, output_cost_per_token=3.0e-6)
+register_model_info('accounts/fireworks/models/kimi-k2p6', vendor_name='fireworks_ai', base='accounts/fireworks/models/kimi-k2p5',
+    supports_reasoning=True, supports_vision=True,
+    input_cost_per_token=0.95e-6, cache_read_input_token_cost=0.16e-6, output_cost_per_token=4.0e-6)
 
 # %% ../nbs/00_types.ipynb #948d55d0
 deepseek_v4_common = dict(**modern_llm, supports_assistant_prefill=True,
@@ -363,6 +365,15 @@ register_model_info('mimo-v2.5-pro', vendor_name='mimo', **mimo_v25_common, base
 register_model_info('mimo-v2.5', vendor_name='mimo', **mimo_v25_common, base='deepseek-v4-pro', base_vendor_name='deepseek',
     input_cost_per_token=0.14e-6,  output_cost_per_token=0.28e-6, cache_read_input_token_cost=0.0028e-6, search_context_cost_per_query=0.005,
     supports_vision=True, supports_image_input=True)
+register_model_info('mimo-v2.5-pro-ultraspeed', vendor_name='mimo', **mimo_v25_common, base='deepseek-v4-pro', base_vendor_name='deepseek',
+    input_cost_per_token=1.305e-6, output_cost_per_token=2.61e-6, cache_read_input_token_cost=0.0108e-6, search_context_cost_per_query=0.005)
+
+# %% ../nbs/00_types.ipynb #defb1c5c
+register_model_info('MiniMax-M3', vendor_name='minimax', **modern_llm, max_input_tokens=512_000, max_output_tokens=512_000, max_tokens=512_000, input_cost_per_token=0.3e-6,  output_cost_per_token=1.2e-6, cache_read_input_token_cost=0.06e-6, supports_vision=True, supports_video_input=True)
+
+register_model_info('claude-fable-5', vendor_name='anthropic', base="claude-opus-4-8",
+        input_cost_per_token=10e-6, cache_creation_input_token_cost=12.5e-6, output_cost_per_token=50e-6,
+        cache_read_input_token_cost=1e-6, search_context_cost_per_query=0.005)
 
 # %% ../nbs/00_types.ipynb #2c23d11e
 codex_pricing = dict(
@@ -388,7 +399,7 @@ def approx_pricing(nm, vendor_name, out=10, cache=80, inp=10, markup=0):
     p = get_model_pricing(nm, vendor_name)
     ic = p.get('cache_creation_input_token_cost', p['input_cost_per_token'])
     res = (p['output_cost_per_token']*out + p['cache_read_input_token_cost']*cache + ic*inp) / (out+cache+inp)
-    if nm in ('claude-opus-4-7','claude-opus-4-8'): res *= 1.5
+    if nm in ('claude-opus-4-7','claude-opus-4-8','claude-fable-5'): res *= 1.5
     return res*(1+markup)
 
 # %% ../nbs/00_types.ipynb #8bfca02d

{python_fastllm-0.0.16 → python_fastllm-0.0.18}/python_fastllm.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: python-fastllm
-Version: 0.0.16
+Version: 0.0.18
 Author-email: Kerem Turgutlu <keremturgutlu@gmail.com>
 License: Apache-2.0
 Project-URL: Repository, https://github.com/AnswerDotAI/fastllm

python_fastllm-0.0.16/fastllm/__init__.py

@@ -1 +0,0 @@
-__version__ = "0.0.16"