v6.9.1 (2026-05-07)

Refactors

• simplify Baidu Qianfan model check and extract model family helpers (d97b1a7, SmallMain)

v6.9.0 (2026-05-02)

Features

• add OpenCode Go providers to README (bffe64e, SmallMain)
• add OpenCode Go providers to well-known provider list (598eac4, Japsterr)

Fixes

• codex: reduce GPT-5.5 max input tokens to 272000 (6f49695, SmallMain)
• codex: prune orphan function calls after compact (a934b1e, moss)
• codex: improve reasoning summary display (bcc706b, moss)
• update model list for Alibaba Cloud Model Studio providers (b82891c, SmallMain)
• detect bad default revision error in commit history check (b8d5c96, SmallMain)

v6.8.5 (2026-04-28)

Fixes

• context: report output buffer usage (5d4cfdf, moss)

v6.8.4 (2026-04-27)

Features

• increase GPT-5.5 max input tokens to 1050000 (933cce8, SmallMain)

Refactors

• increase default timeouts and format code in utils (c3e0bb5, SmallMain)

v6.8.3 (2026-04-25)

Refactors

• simplify DeepSeek V4 and GLM feature checkers (5a3aa3d, SmallMain)

v6.8.2 (2026-04-24)

Features

• increase GPT-5.5 max input tokens to 400000 (83cff25, SmallMain)

v6.8.1 (2026-04-24)

Refactors

• update Codex client with new UA and image generation tool (b9f620d, SmallMain)

v6.8.0 (2026-04-24)

Features

• add GPT-5.5 model support (33aa497, SmallMain)

v6.7.1 (2026-04-24)

Fixes

• openai: handle undefined/null reasoning_content properly (ab01200, SmallMain)

v6.7.0 (2026-04-24)

Features

• set DeepSeek V4 reasoning effort to max (bf62018, SmallMain)
• add DeepSeek V4 support with reasoning_effort param (6272955, SmallMain)