From 8ead508596929869e0ea9e38b1cf1f2993722ef5 Mon Sep 17 00:00:00 2001
From: nuomi1 <nuomi1@qq.com>
Date: Wed, 8 Jul 2026 13:17:49 +0800
Subject: [PATCH 1/2] feat(models): add SenseNova-6.7-Flash-Lite and
 SenseNova-U1-Fast metadata

Add model metadata entries for two new SenseNova models:
- SenseNova-6.7-Flash-Lite: lightweight multimodal agent model with image input support
- SenseNova-U1-Fast: accelerated version for infographics generation with image output
---
 models/sensenova/sensenova-6.7-flash-lite.toml | 18 ++++++++++++++++++
 models/sensenova/sensenova-u1-fast.toml        | 18 ++++++++++++++++++
 2 files changed, 36 insertions(+)
 create mode 100644 models/sensenova/sensenova-6.7-flash-lite.toml
 create mode 100644 models/sensenova/sensenova-u1-fast.toml

diff --git a/models/sensenova/sensenova-6.7-flash-lite.toml b/models/sensenova/sensenova-6.7-flash-lite.toml
new file mode 100644
index 0000000000..7167589bac
--- /dev/null
+++ b/models/sensenova/sensenova-6.7-flash-lite.toml
@@ -0,0 +1,18 @@
+name = "SenseNova-6.7-Flash-Lite"
+description = "Lightweight multimodal agent model designed for real-world workflows, supporting text conversation and image input understanding"
+release_date = "2026-05-08"
+last_updated = "2026-05-08"
+attachment = true
+reasoning = true
+temperature = true
+tool_call = true
+structured_output = true
+open_weights = false
+
+[limit]
+context = 262_144
+output = 65_536
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/models/sensenova/sensenova-u1-fast.toml b/models/sensenova/sensenova-u1-fast.toml
new file mode 100644
index 0000000000..9863035100
--- /dev/null
+++ b/models/sensenova/sensenova-u1-fast.toml
@@ -0,0 +1,18 @@
+name = "SenseNova-U1-Fast"
+description = "Accelerated version based on SenseNova U1, specialized for infographics generation"
+release_date = "2026-05-08"
+last_updated = "2026-05-08"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+structured_output = true
+open_weights = false
+
+[limit]
+context = 262_144
+output = 65_536
+
+[modalities]
+input = ["text"]
+output = ["image"]

From 0b3e9e2c7d97877f11342d8a7abca73c4dd192e0 Mon Sep 17 00:00:00 2001
From: nuomi1 <nuomi1@qq.com>
Date: Wed, 8 Jul 2026 18:46:46 +0800
Subject: [PATCH 2/2] feat(providers/sensenova): add SenseNova provider with
 initial models

Add SenseNova provider configuration and three models:
- deepseek-v4-flash (interleaved reasoning with effort control)
- sensenova-6.7-flash-lite (interleaved reasoning with effort control)
- sensenova-u1-fast (no reasoning support)
---
 providers/sensenova/models/deepseek-v4-flash.toml    | 12 ++++++++++++
 .../sensenova/models/sensenova-6.7-flash-lite.toml   | 12 ++++++++++++
 providers/sensenova/models/sensenova-u1-fast.toml    |  7 +++++++
 providers/sensenova/provider.toml                    | 12 ++++++++++++
 4 files changed, 43 insertions(+)
 create mode 100644 providers/sensenova/models/deepseek-v4-flash.toml
 create mode 100644 providers/sensenova/models/sensenova-6.7-flash-lite.toml
 create mode 100644 providers/sensenova/models/sensenova-u1-fast.toml
 create mode 100644 providers/sensenova/provider.toml

diff --git a/providers/sensenova/models/deepseek-v4-flash.toml b/providers/sensenova/models/deepseek-v4-flash.toml
new file mode 100644
index 0000000000..1411da3ce0
--- /dev/null
+++ b/providers/sensenova/models/deepseek-v4-flash.toml
@@ -0,0 +1,12 @@
+base_model = "deepseek/deepseek-v4-flash"
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[interleaved]
+field = "reasoning_content"
+
+[cost]
+input = 0
+output = 0
diff --git a/providers/sensenova/models/sensenova-6.7-flash-lite.toml b/providers/sensenova/models/sensenova-6.7-flash-lite.toml
new file mode 100644
index 0000000000..9ae808d5c9
--- /dev/null
+++ b/providers/sensenova/models/sensenova-6.7-flash-lite.toml
@@ -0,0 +1,12 @@
+base_model = "sensenova/sensenova-6.7-flash-lite"
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[interleaved]
+field = "reasoning_content"
+
+[cost]
+input = 0
+output = 0
diff --git a/providers/sensenova/models/sensenova-u1-fast.toml b/providers/sensenova/models/sensenova-u1-fast.toml
new file mode 100644
index 0000000000..e7a81d7b31
--- /dev/null
+++ b/providers/sensenova/models/sensenova-u1-fast.toml
@@ -0,0 +1,7 @@
+base_model = "sensenova/sensenova-u1-fast"
+
+reasoning_options = []
+
+[cost]
+input = 0
+output = 0
diff --git a/providers/sensenova/provider.toml b/providers/sensenova/provider.toml
new file mode 100644
index 0000000000..166bd9b933
--- /dev/null
+++ b/providers/sensenova/provider.toml
@@ -0,0 +1,12 @@
+# SenseNova reasoning controls (accessed 2026-07-08):
+# The API at https://token.sensenova.cn/v1 is OpenAI-compatible. Reasoning
+# effort (low/medium/high) is supported via reasoning_effort on models that
+# declare reasoning_options, with interleaved reasoning returned in the
+# reasoning_content field. sensenova-u1-fast does not support reasoning.
+# Source:
+# https://platform.sensenova.cn/en/docs
+name = "SenseNova"
+env = ["SENSENOVA_API_KEY"]
+npm = "@ai-sdk/openai-compatible"
+api = "https://token.sensenova.cn/v1"
+doc = "https://platform.sensenova.cn/en/docs"