feat: add novita client (#1132)

4 files changed, 65 insertions, 15 deletions

diff --git a/Argcfile.sh b/Argcfile.sh
index ea77211..4ef3d16 100755
--- a/Argcfile.sh
+++ b/Argcfile.sh
@@ -311,6 +311,7 @@ _argc_before() {
         minimax,MiniMax-Text-01,https://api.minimax.chat/v1 \
         mistral,mistral-small-latest,https://api.mistral.ai/v1 \
         moonshot,moonshot-v1-8k,https://api.moonshot.cn/v1 \
+        novita,meta-llama/llama-3.1-8b-instruct,https://api.novita.ai/v3/openai \
         openrouter,openai/gpt-4o-mini,https://openrouter.ai/api/v1 \
         ollama,llama3.1:latest,http://${OLLAMA_HOST:-"127.0.0.1:11434"}/v1 \
         perplexity,llama-3.1-8b-instruct,https://api.perplexity.ai \
diff --git a/config.example.yaml b/config.example.yaml
index 8e81f8a..ede0fe7 100644
--- a/config.example.yaml
+++ b/config.example.yaml
@@ -334,6 +334,12 @@ clients:
     name: hyperbolic
     api_base: https://api.hyperbolic.xyz/v1
     api_key: xxx
+  
+  # See https://novita.ai/docs/guides/introduction
+  - type: openai-compatible
+    name: novita
+    api_base: https://api.novita.ai/v3/openai
+    api_key: xxx
 
   # ----- RAG dedicated -----
 
diff --git a/models.yaml b/models.yaml
index 4816586..3a39c9e 100644
--- a/models.yaml
+++ b/models.yaml
@@ -394,6 +394,10 @@
       input_price: 0
       output_price: 0
       supports_vision: true
+    - name: deepseek-r1-distill-llama-70b
+      max_input_tokens: 131072
+      input_price: 0
+      output_price: 0
 
 # Links:
 #  - https://ollama.com/library
@@ -417,8 +421,6 @@
       supports_function_calling: true
     - name: deepseek-r1
       max_input_tokens: 131072
-    - name: phi4
-      max_input_tokens: 16384
     - name: nomic-embed-text
       type: embedding
       max_tokens_per_chunk: 8192
@@ -1265,14 +1267,10 @@
       max_input_tokens: 163840
       input_price: 0.55
       output_price: 2.19
-    - name: microsoft/phi-4
-      max_input_tokens: 16384
-      input_price: 0.07
-      output_price: 0.14
-    - name: microsoft/phi-3.5-mini-128k-instruct
-      max_input_tokens: 128000
-      input_price: 0.1
-      output_price: 0.1
+    - name: deepseek/deepseek-r1-distill-llama-70b
+      max_input_tokens: 131072
+      input_price: 0.23
+      output_price: 0.69
     - name: qwen/qwen-2.5-72b-instruct
       max_input_tokens: 131072
       input_price: 0.35
@@ -1482,10 +1480,10 @@
       max_input_tokens: 16000
       input_price: 0.85
       output_price: 2.5
-    - name: microsoft/phi-4
-      max_input_tokens: 16384
-      input_price: 0.07
-      output_price: 0.14
+    - name: deepseek-ai/DeepSeek-R1-Distill-Llama-70B
+      max_input_tokens: 131072
+      input_price: 0.23
+      output_price: 0.69
     - name: BAAI/bge-large-en-v1.5
       type: embedding
       input_price: 0.01
@@ -1819,6 +1817,50 @@
       supports_vision: true
 
 # Links:
+#  - https://novita.ai/models
+#  - https://novita.ai/docs/guides/llm-api
+- provider: novita
+  models:
+    - name: meta-llama/llama-3.3-70b-instruct
+      max_input_tokens: 131072
+      input_price: 0.39
+      output_price: 0.39
+    - name: meta-llama/llama-3.1-70b-instruct
+      max_input_tokens: 32768
+      input_price: 0.34
+      output_price: 0.39
+    - name: meta-llama/llama-3.1-8b-instruct
+      max_input_tokens: 16384
+      input_price: 0.05
+      output_price: 0.05
+    - name: meta-llama/llama-3.2-11b-vision-instruct
+      max_input_tokens: 32768
+      input_price: 0.06
+      output_price: 0.06
+      supports_vision: true
+    - name: qwen/qwen-2.5-72b-instruct
+      max_input_tokens: 32000
+      input_price: 0.38
+      output_price: 0.4
+    - name: qwen/qwen-2-vl-72b-instruct
+      max_input_tokens: 32768
+      input_price: 0.45
+      output_price: 0.45
+      supports_vision: true
+    - name: deepseek/deepseek_v3
+      max_input_tokens: 64000
+      input_price: 0.89
+      output_price: 0.89
+    - name: deepseek/deepseek-r1
+      max_input_tokens: 64000
+      input_price: 4
+      output_price: 4
+    - name: deepseek/deepseek-r1-distill-llama-70b
+      max_input_tokens: 32000
+      input_price: 0.8
+      output_price: 0.8
+
+# Links:
 #  - https://jina.ai/models
 #  - https://api.jina.ai/redoc
 - provider: jina
diff --git a/src/client/mod.rs b/src/client/mod.rs
index c87d2b5..c6b65bc 100644
--- a/src/client/mod.rs
+++ b/src/client/mod.rs
@@ -33,7 +33,7 @@ register_client!(
     (bedrock, "bedrock", BedrockConfig, BedrockClient),
 );
 
-pub const OPENAI_COMPATIBLE_PROVIDERS: [(&str, &str); 24] = [
+pub const OPENAI_COMPATIBLE_PROVIDERS: [(&str, &str); 25] = [
     ("ai21", "https://api.ai21.com/studio/v1"),
     (
         "cloudflare",
@@ -51,6 +51,7 @@ pub const OPENAI_COMPATIBLE_PROVIDERS: [(&str, &str); 24] = [
     ("minimax", "https://api.minimax.chat/v1"),
     ("mistral", "https://api.mistral.ai/v1"),
     ("moonshot", "https://api.moonshot.cn/v1"),
+    ("novita", "https://api.novita.ai/v3/openai"),
     ("openrouter", "https://openrouter.ai/api/v1"),
     ("ollama", "http://{OLLAMA_HOST}:11434/v1"),
     ("perplexity", "https://api.perplexity.ai"),