omk-ai 1.5.7 → 1.5.9

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/CHANGELOG.md CHANGED
@@ -2,6 +2,24 @@
2
2
 
3
3
  ## [Unreleased]
4
4
 
5
+ ## [1.5.9] - 2026-10-26
6
+
7
+ ### Added
8
+
9
+ - Kimi K3.1 Pro: `k3.1-pro` and `k3.1-pro-highspeed` in the `kimi-coding` catalog (thinking always on, fixed sampling) and `kimi-k3.1-pro` and `kimi-k3.1-pro-highspeed` in the `moonshotai` catalog with published platform pricing (#280).
10
+
11
+ ### Changed
12
+
13
+ - Regenerated the model catalog from models.dev and live provider listings (2026-10-26). Models that upstream marked deprecated or delisted are removed, including GitHub Copilot `gpt-5.4`, `gpt-5.5` and `gpt-5-mini`, older Mistral open models, Together `MiniMax-M2.7` and Xiaomi `mimo-v2` models. OpenCode Go replaces `gemini-4-argon-contributor` with `gemini-4-argon` and adds `kimi-k3.1-pro` and `minimax-m3.1-flash`. Direct Mistral Large 4 pricing now uses the list price after the launch sale ended.
14
+
15
+ ### Fixed
16
+
17
+ - WorkBuddy refusals with a `{ code, msg, requestId }` body now preserve the provider's message, code and request ID instead of reporting `<status> status code (no body)`. Normalization is scoped to WorkBuddy; HTTP status, headers, successful streams, OpenAI-shaped errors and empty bodies retain their existing behavior.
18
+ - Recognize OMK's local safe-context-window capacity rejection as context overflow, allowing session classification and bounded compaction recovery to share the provider-overflow path.
19
+
20
+ ## [1.5.8] - 2026-10-24
21
+
22
+
5
23
  ## [1.5.7] - 2026-10-23
6
24
 
7
25