diff --git a/providers/regolo-ai/logo.svg b/providers/regolo-ai/logo.svg
new file mode 100644
index 000000000..320803a36
--- /dev/null
+++ b/providers/regolo-ai/logo.svg
@@ -0,0 +1,46 @@
+<?xml version="1.0" encoding="UTF-8" standalone="no"?>
+<!-- Created with Inkscape (http://www.inkscape.org/) -->
+
+<svg
+   width="32"
+   height="32"
+   viewBox="0 0 32 32"
+   version="1.1"
+   id="svg1"
+   xmlns:inkscape="http://www.inkscape.org/namespaces/inkscape"
+   xmlns:sodipodi="http://sodipodi.sourceforge.net/DTD/sodipodi-0.dtd"
+   xmlns="http://www.w3.org/2000/svg"
+   xmlns:svg="http://www.w3.org/2000/svg">
+  <sodipodi:namedview
+     id="namedview1"
+     pagecolor="#ffffff"
+     bordercolor="#000000"
+     borderopacity="0.25"
+     inkscape:showpageshadow="2"
+     inkscape:pageopacity="0.0"
+     inkscape:pagecheckerboard="0"
+     inkscape:deskcolor="#d1d1d1"
+     inkscape:document-units="px" />
+  <defs
+     id="defs1" />
+  <g
+     inkscape:label="Livello 1"
+     inkscape:groupmode="layer"
+     id="layer1">
+    <path
+       class="cls-1"
+       d="m 30.997126,19.704021 v 10.396552 c 0,0.09195 -0.01436,0.178162 -0.04023,0.261494 -0.112068,0.364944 -0.454022,0.632184 -0.859194,0.632184 H 19.70115 c -0.497128,0 -0.899426,-0.405172 -0.899426,-0.899424 v -8.252874 c 0,-0.0431 0,-0.08333 0,-0.126436 v -2.011496 c 0,-0.497126 0.402298,-0.899424 0.899426,-0.899424 h 10.399424 c 0.431036,0 0.795978,0.30747 0.879312,0.71839 0.01436,0.06035 0.02011,0.117816 0.02011,0.181034 z"
+       id="path1"
+       style="fill:#303030;stroke-width:0.287356" />
+    <path
+       class="cls-1"
+       d="m 30.997126,19.704021 v 10.396552 c 0,0.09195 -0.01436,0.178162 -0.04023,0.261494 -0.112068,0.364944 -0.454022,0.632184 -0.859194,0.632184 H 19.70115 c -0.497128,0 -0.899426,-0.405172 -0.899426,-0.899424 v -8.252874 c 0,-0.0431 0,-0.08333 0,-0.126436 v -2.011496 c 0,-0.497126 0.402298,-0.899424 0.899426,-0.899424 h 10.399424 c 0.431036,0 0.795978,0.30747 0.879312,0.71839 0.01436,0.06035 0.02011,0.117816 0.02011,0.181034 z"
+       id="path2"
+       style="fill:#4d4d4d;stroke-width:0.287356" />
+    <path
+       class="cls-2"
+       d="M 30.100574,1.002874 H 19.704022 c -0.497126,0 -0.899424,0.402298 -0.899424,0.899425 v 8.25 c 0,1.353448 -0.54023,2.652298 -1.494254,3.609194 l -3.545976,3.543104 c -0.954024,0.956896 -2.255748,1.494252 -3.606322,1.494252 H 6.454023 c -0.635058,0 -1.146552,-0.514366 -1.146552,-1.14655 v -3.31322 c 0,-0.635056 0.514368,-1.146551 1.146552,-1.146551 h 5.844827 c 0.405172,0 0.744254,-0.267241 0.859196,-0.632184 0.02586,-0.08333 0.04023,-0.16954 0.04023,-0.261494 V 1.899425 c 0,-0.06322 -0.0057,-0.120689 -0.02012,-0.181034 C 13.094826,1.307471 12.729886,1 12.29885,1 H 1.899425 C 1.402299,1 1,1.402299 1,1.899425 V 3.91092 c 0,0.0431 0,0.08333 0,0.126436 v 8.252873 c 0,0.494253 0.405172,0.899426 0.899425,0.899426 h 0.908046 c 0.635058,0 1.146552,0.514368 1.146552,1.146552 v 3.313218 c 0,0.635058 -0.514368,1.146552 -1.146552,1.146552 H 1.905172 c -0.497126,0 -0.899425,0.402298 -0.899425,0.899424 v 10.399426 c 0,0.494252 0.405173,0.899424 0.899425,0.899424 h 10.399426 c 0.497126,0 0.899424,-0.405172 0.899424,-0.899424 v -8.252874 c 0,-1.350574 0.537358,-2.6523 1.494254,-3.606322 l 3.543104,-3.545976 c 0.956896,-0.954024 2.255746,-1.494253 3.609194,-1.494253 h 8.25 C 30.597702,13.195402 31,12.793103 31,12.295977 V 1.899425 C 31,1.402299 30.597702,1 30.100574,1 Z"
+       id="path3"
+       style="fill:#000000;stroke-width:0.287356" />
+  </g>
+</svg>
diff --git a/providers/regolo-ai/models/Llama-3.1-8B-Instruct.toml b/providers/regolo-ai/models/Llama-3.1-8B-Instruct.toml
new file mode 100644
index 000000000..e75d7de2e
--- /dev/null
+++ b/providers/regolo-ai/models/Llama-3.1-8B-Instruct.toml
@@ -0,0 +1,24 @@
+name = "Llama‑3.1‑8B‑Instruct"
+family = "llama-3.1"
+release_date = "2025-04-07"
+last_updated = "2025-04-07"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = false
+
+[description]
+text = "Llama‑3.1‑8B‑Instruct is a compact multilingual language model by Meta, designed for instruction following and conversational use, with reliable performance in text understanding, code generation, and structured outputs."
+
+[cost]
+input = 0.05   # € per 1M tokens (0.00000005 * 1_000_000)
+output = 0.25  # € per 1M tokens (0.00000025 * 1_000_000)
+
+[limit]
+context = 120000
+output = 120000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/regolo-ai/models/Llama-3.3-70B-Instruct.toml b/providers/regolo-ai/models/Llama-3.3-70B-Instruct.toml
new file mode 100644
index 000000000..5a8a12c0b
--- /dev/null
+++ b/providers/regolo-ai/models/Llama-3.3-70B-Instruct.toml
@@ -0,0 +1,24 @@
+name = "Llama‑3.3‑70B‑Instruct"
+family = "llama-3.3"
+release_date = "2025-04-28"
+last_updated = "2025-04-28"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = false
+
+[description]
+text = "Llama‑3.3‑70B‑Instruct is a 70B‑parameter multilingual language model by Meta, designed for instruction following, long‑context understanding, and structured reasoning, with strong capabilities in code generation and complex analytical tasks."
+
+[cost]
+input = 0.60    # € per 1M tokens (0.0000006 * 1_000_000)
+output = 2.70   # € per 1M tokens (0.0000027 * 1_000_000)
+
+[limit]
+context = 16000
+output = 16000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/regolo-ai/models/Qwen3-8B.toml b/providers/regolo-ai/models/Qwen3-8B.toml
new file mode 100644
index 000000000..cb4658b8c
--- /dev/null
+++ b/providers/regolo-ai/models/Qwen3-8B.toml
@@ -0,0 +1,24 @@
+name = "Qwen3‑8B"
+family = "qwen3-8b"
+release_date = "2025-04-28"
+last_updated = "2025-04-28"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+open_weights = false
+
+[description]
+text = "Qwen3‑8B is a versatile 8.2‑billion parameter language model that uniquely allows switching between a step‑by‑step 'thinking mode' for complex tasks and a fast 'non‑thinking mode' for general conversation."
+
+[cost]
+input = 0.07   # € per 1M tokens (placeholder from Regolo pricing)
+    output = 0.35   # € per 1M tokens (placeholder from Regolo pricing)
+
+[limit]
+context = 40960
+output = 40960
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/regolo-ai/models/deepseek-r1-70b.toml b/providers/regolo-ai/models/deepseek-r1-70b.toml
new file mode 100644
index 000000000..8565a6674
--- /dev/null
+++ b/providers/regolo-ai/models/deepseek-r1-70b.toml
@@ -0,0 +1,24 @@
+name = "deepseek‑r1‑70b"
+family = "deepseek-r1"
+release_date = "2025-01-20"
+last_updated = "2025-01-20"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+open_weights = false
+
+[description]
+text = "DeepSeek‑R1‑Distill‑Llama‑70B is a 70B‑parameter distilled LLM, combining reasoning, speed, and accuracy for code, math, and complex logic tasks."
+
+[cost]
+input = 0.60    # € per 1M tokens (0.0000006 * 1_000_000)
+output = 2.70   # € per 1M tokens (0.0000027 * 1_000_000)
+
+[limit]
+context = 128000
+output = 128000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/regolo-ai/models/gpt-oss-120b.toml b/providers/regolo-ai/models/gpt-oss-120b.toml
new file mode 100644
index 000000000..debc1d613
--- /dev/null
+++ b/providers/regolo-ai/models/gpt-oss-120b.toml
@@ -0,0 +1,24 @@
+name = "gpt-oss-120b"
+family = "gpt-oss"
+release_date = "2025-08-05"
+last_updated = "2025-08-05"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+open_weights = false
+
+[description]
+text = "GPT‑OSS‑120B is an open‑weight 117B‑parameter Mixture‑of‑Experts model by OpenAI, using only 5.1B active parameters per token. Supports reasoning, chain‑of‑thought, tool use, and fine‑tuning."
+
+[cost]
+input = 1.00    # € per 1M tokens (0.000001 * 1_000_000)
+output = 4.20   # € per 1M tokens (0.0000042 * 1_000_000)
+
+[limit]
+context = 128000
+output = 128000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/regolo-ai/models/mistral-small3.2.toml b/providers/regolo-ai/models/mistral-small3.2.toml
new file mode 100644
index 000000000..cdef2e083
--- /dev/null
+++ b/providers/regolo-ai/models/mistral-small3.2.toml
@@ -0,0 +1,24 @@
+name = "mistral‑small3.2"
+family = "mistral-small3.2"
+release_date = "2025-01-31"
+last_updated = "2025-01-31"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+open_weights = false
+
+[description]
+text = "Mistral‑Small‑3.2‑24B‑Instruct‑2506: 24B multimodal instruction‑tuned model, optimized for reasoning and STEM, supports robust function calling, reduces repetition, handles both text and vision inputs efficiently."
+
+[cost]
+input = 0.50    # € per 1M tokens (0.0000005 * 1_000_000)
+output = 2.20   # € per 1M tokens (0.0000022 * 1_000_000)
+
+[limit]
+context = 120000
+output = 120000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/regolo-ai/models/qwen3-coder-30b.toml b/providers/regolo-ai/models/qwen3-coder-30b.toml
new file mode 100644
index 000000000..7db53715b
--- /dev/null
+++ b/providers/regolo-ai/models/qwen3-coder-30b.toml
@@ -0,0 +1,24 @@
+name = "qwen3‑coder‑30b"
+family = "qwen3-coder-30b"
+release_date = "2025-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = false
+
+[description]
+text = "Qwen3‑Coder‑30B is a large‑scale MOE coding model designed for complex software tasks, featuring strong code understanding, long‑context reasoning, and reliable structured outputs for repository‑level analysis and agentic programming workflows."
+
+[cost]
+input = 0.50    # € per 1M tokens (0.0000005 * 1_000_000)
+output = 2.00   # € per 1M tokens (0.0000020 * 1_000_000)
+
+[limit]
+context = 120000
+output = 120000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/regolo-ai/models/qwen3-coder-next.toml b/providers/regolo-ai/models/qwen3-coder-next.toml
new file mode 100644
index 000000000..4fcbf2b9d
--- /dev/null
+++ b/providers/regolo-ai/models/qwen3-coder-next.toml
@@ -0,0 +1,24 @@
+name = "qwen3-coder-next"
+family = "qwen3-coder-next-fp8"
+release_date = "2025-02-05"
+last_updated = "2025-02-05"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = false
+
+[description]
+text = "Qwen3-Coder-Next-FP8 is a 3B activated (80B total) MoE coding model designed for coding agents and local development. Features 256K context length and advanced tool calling capabilities."
+
+[cost]
+input = 0.50    # € per 1M tokens
+output = 2.00   # € per 1M tokens
+
+[limit]
+context = 262144
+output = 65536
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/regolo-ai/models/qwen3-vl-32b.toml b/providers/regolo-ai/models/qwen3-vl-32b.toml
new file mode 100644
index 000000000..f09f4d1e8
--- /dev/null
+++ b/providers/regolo-ai/models/qwen3-vl-32b.toml
@@ -0,0 +1,26 @@
+name = "qwen3‑vl‑32b"
+family = "qwen3-vl-32b"
+release_date = "2025-01-01"
+last_updated = "2025-01-01"
+attachment = true
+reasoning = true
+temperature = true
+tool_call = true
+open_weights = false
+
+[description]
+text = "Qwen3‑VL‑32B is a 32‑billion‑parameter multimodal vision‑language model from the Qwen 3 family, supporting text, image, audio and video inputs."
+
+[cost]
+# Prices need verification against Regolo pricing page
+input = 0.50   # € per 1M tokens (official price)
+output = 2.50  # € per 1M tokens (official price)
+
+[limit]
+context = 128000
+output = 128000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
+
diff --git a/providers/regolo-ai/provider.toml b/providers/regolo-ai/provider.toml
new file mode 100644
index 000000000..c0aab7eb4
--- /dev/null
+++ b/providers/regolo-ai/provider.toml
@@ -0,0 +1,5 @@
+name = "Regolo AI"
+env = ["REGOLO_API_KEY"]
+npm = "@ai-sdk/openai-compatible"
+doc = "https://docs.regolo.ai/"
+api = "https://api.regolo.ai/v1"