From ceb91b48dfb6f0aff915a9e1fa0954ef3fd75b1f Mon Sep 17 00:00:00 2001
From: lcwecker <werkerrrrr@gmail.com>
Date: Sun, 1 Mar 2026 20:10:33 +0800
Subject: [PATCH] Add three new NVIDIA NIM models: Qwen3.5-397B, MiniMax-M2.5,
 and Step-3.5-Flash

- Add qwen3.5-397b-a17b: 397B MoE multimodal model with 262K context (NVIDIA NIM)
- Add minimax-m2.5: 230B coding and reasoning model with 204K context (NVIDIA NIM)
- Add step-3.5-flash: 196B MoE reasoning model with 256K context (NVIDIA NIM)
- Create stepfun-ai provider directory for Step-3.5-Flash
---
 .../nvidia/models/minimaxai/minimax-m2.5.toml | 21 +++++++++++++++++++
 .../nvidia/models/qwen/qwen3.5-397b-a17b.toml | 21 +++++++++++++++++++
 .../models/stepfun-ai/step-3.5-flash.toml     | 21 +++++++++++++++++++
 3 files changed, 63 insertions(+)
 create mode 100644 providers/nvidia/models/minimaxai/minimax-m2.5.toml
 create mode 100644 providers/nvidia/models/qwen/qwen3.5-397b-a17b.toml
 create mode 100644 providers/nvidia/models/stepfun-ai/step-3.5-flash.toml

diff --git a/providers/nvidia/models/minimaxai/minimax-m2.5.toml b/providers/nvidia/models/minimaxai/minimax-m2.5.toml
new file mode 100644
index 000000000..6ad07e085
--- /dev/null
+++ b/providers/nvidia/models/minimaxai/minimax-m2.5.toml
@@ -0,0 +1,21 @@
+name = "MiniMax-M2.5"
+family = "minimax"
+release_date = "2026-02-12"
+last_updated = "2026-02-26"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 204_800
+output = 131_072
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/nvidia/models/qwen/qwen3.5-397b-a17b.toml b/providers/nvidia/models/qwen/qwen3.5-397b-a17b.toml
new file mode 100644
index 000000000..3408f91f2
--- /dev/null
+++ b/providers/nvidia/models/qwen/qwen3.5-397b-a17b.toml
@@ -0,0 +1,21 @@
+name = "Qwen3.5-397B-A17B"
+family = "qwen"
+release_date = "2026-02-16"
+last_updated = "2026-02-16"
+attachment = true
+reasoning = true
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 262_144
+output = 81_920
+
+[modalities]
+input = ["text", "image", "video"]
+output = ["text"]
diff --git a/providers/nvidia/models/stepfun-ai/step-3.5-flash.toml b/providers/nvidia/models/stepfun-ai/step-3.5-flash.toml
new file mode 100644
index 000000000..cbe839bb4
--- /dev/null
+++ b/providers/nvidia/models/stepfun-ai/step-3.5-flash.toml
@@ -0,0 +1,21 @@
+name = "Step-3.5-Flash"
+family = "step"
+release_date = "2026-02-01"
+last_updated = "2026-02-01"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.0
+output = 0.0
+
+[limit]
+context = 256_000
+output = 32_768
+
+[modalities]
+input = ["text"]
+output = ["text"]