From 7db78fc61ae67b9ef554cd5d5f21191aaee9095b Mon Sep 17 00:00:00 2001
From: Tai An <antai12232931@outlook.com>
Date: Mon, 4 May 2026 03:22:38 -0700
Subject: [PATCH] feat(model_prices): add ai21 jamba-mini-2 and dated
 jamba-large-1.7 aliases (#27094)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

AI21 has rolled out a new generation. Per the AI21 API ( /studio/v1/models )
and pricing page, the active direct-API models are now:

  - jamba-large-1.7-2025-07  (alias: jamba-large-1.7)  - already present
  - jamba-mini-2-2026-01     (alias: jamba-mini-2)     - missing

This change adds:
  * jamba-large-1.7-2025-07  (dated alias of jamba-large-1.7,
                              already present at $2 / $8 per 1M)
  * jamba-mini-2             ($0.2 / $0.4 per 1M, 256K context)
  * jamba-mini-2-2026-01     (dated form of jamba-mini-2)

Pricing matches the AI21 public pricing page (Jamba Mini = $0.2/1M input,
$0.4/1M output; Jamba Large = $2/1M input, $8/1M output) and is consistent
with the existing jamba-large-1.7 / jamba-mini-1.7 entries. Context window
follows the existing 256K Jamba family setting.

Refs #27094.

The legacy J2 entries (j2-light, j2-mid, j2-ultra) are intentionally left
in place so cost lookups for historical request logs still resolve - the
litellm convention elsewhere is to keep retired model entries for backward
compatibility rather than delete them.

The credential-binding UI bug also mentioned in #27094 is out of scope
for this data-only patch and should be tracked separately.

🤖 Generated with [Claude Code](https://claude.com/claude-code)
---
 ...odel_prices_and_context_window_backup.json | 30 +++++++++++++++++++
 model_prices_and_context_window.json          | 30 +++++++++++++++++++
 2 files changed, 60 insertions(+)

diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json
index 13a45fd1650..6877f42152d 100644
--- a/litellm/model_prices_and_context_window_backup.json
+++ b/litellm/model_prices_and_context_window_backup.json
@@ -21977,6 +21977,16 @@
         "output_cost_per_token": 8e-06,
         "supports_tool_choice": true
     },
+    "jamba-large-1.7-2025-07": {
+        "input_cost_per_token": 2e-06,
+        "litellm_provider": "ai21",
+        "max_input_tokens": 256000,
+        "max_output_tokens": 256000,
+        "max_tokens": 256000,
+        "mode": "chat",
+        "output_cost_per_token": 8e-06,
+        "supports_tool_choice": true
+    },
     "jamba-mini-1.6": {
         "input_cost_per_token": 2e-07,
         "litellm_provider": "ai21",
@@ -21997,6 +22007,26 @@
         "output_cost_per_token": 4e-07,
         "supports_tool_choice": true
     },
+    "jamba-mini-2": {
+        "input_cost_per_token": 2e-07,
+        "litellm_provider": "ai21",
+        "max_input_tokens": 256000,
+        "max_output_tokens": 256000,
+        "max_tokens": 256000,
+        "mode": "chat",
+        "output_cost_per_token": 4e-07,
+        "supports_tool_choice": true
+    },
+    "jamba-mini-2-2026-01": {
+        "input_cost_per_token": 2e-07,
+        "litellm_provider": "ai21",
+        "max_input_tokens": 256000,
+        "max_output_tokens": 256000,
+        "max_tokens": 256000,
+        "mode": "chat",
+        "output_cost_per_token": 4e-07,
+        "supports_tool_choice": true
+    },
     "jina-reranker-v2-base-multilingual": {
         "input_cost_per_token": 1.8e-08,
         "litellm_provider": "jina_ai",
diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json
index bbe13442d63..e9130efab25 100644
--- a/model_prices_and_context_window.json
+++ b/model_prices_and_context_window.json
@@ -22025,6 +22025,16 @@
         "output_cost_per_token": 8e-06,
         "supports_tool_choice": true
     },
+    "jamba-large-1.7-2025-07": {
+        "input_cost_per_token": 2e-06,
+        "litellm_provider": "ai21",
+        "max_input_tokens": 256000,
+        "max_output_tokens": 256000,
+        "max_tokens": 256000,
+        "mode": "chat",
+        "output_cost_per_token": 8e-06,
+        "supports_tool_choice": true
+    },
     "jamba-mini-1.6": {
         "input_cost_per_token": 2e-07,
         "litellm_provider": "ai21",
@@ -22045,6 +22055,26 @@
         "output_cost_per_token": 4e-07,
         "supports_tool_choice": true
     },
+    "jamba-mini-2": {
+        "input_cost_per_token": 2e-07,
+        "litellm_provider": "ai21",
+        "max_input_tokens": 256000,
+        "max_output_tokens": 256000,
+        "max_tokens": 256000,
+        "mode": "chat",
+        "output_cost_per_token": 4e-07,
+        "supports_tool_choice": true
+    },
+    "jamba-mini-2-2026-01": {
+        "input_cost_per_token": 2e-07,
+        "litellm_provider": "ai21",
+        "max_input_tokens": 256000,
+        "max_output_tokens": 256000,
+        "max_tokens": 256000,
+        "mode": "chat",
+        "output_cost_per_token": 4e-07,
+        "supports_tool_choice": true
+    },
     "jina-reranker-v2-base-multilingual": {
         "input_cost_per_token": 1.8e-08,
         "litellm_provider": "jina_ai",