Skip to content

fix: ClinePass thinking-effort levels are wrong for all models: registry shows none/low/medium/high/xhigh, but DeepSeek V4.1 Flash officially supports off/low/high/max #7534

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
github-actions wants to merge 1 commit into
base: dev
Choose a base branch
from
Open

fix: ClinePass thinking-effort levels are wrong for all models: registry shows none/low/medium/high/xhigh, but DeepSeek V4.1 Flash officially supports off/low/high/max #7534

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 5 additions & 1 deletion providers/cline-pass/models/cline-pass/deepseek-v4-flash.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,9 @@
# https://docs.cline.bot/getting-started/clinepass
# https://api-docs.deepseek.com/guides/thinking_mode/ (OpenAI: thinking.type + reasoning_effort low|high|max)
# Effort: reasoning_effort = none|low|high|max (none disables thinking; DeepSeek Flash graded set)
# Legacy Flash slug; lab serves V4.1-Flash. Same reasoning surface as deepseek-v4.1-flash.
base_model = "deepseek/deepseek-v4-flash"
reasoning_options = [{ type = "effort", values = ["none", "low", "medium", "high", "xhigh"] }]
reasoning_options = [{ type = "effort", values = ["none", "low", "high", "max"] }]

[cost]
input = 0.14
Expand Down
6 changes: 5 additions & 1 deletion providers/cline-pass/models/cline-pass/deepseek-v4-pro.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,9 @@
# https://docs.cline.bot/getting-started/clinepass
# https://api-docs.deepseek.com/guides/thinking_mode/ (OpenAI: thinking.type + reasoning_effort)
# First-party DeepSeek Pro: effort high|max (lab providers/deepseek/models/deepseek-v4-pro.toml)
# Effort: reasoning_effort = none|high|max (none disables thinking)
base_model = "deepseek/deepseek-v4-pro"
reasoning_options = [{ type = "effort", values = ["none", "low", "medium", "high", "xhigh"] }]
reasoning_options = [{ type = "effort", values = ["none", "high", "max"] }]

[cost]
input = 1.74
Expand Down
6 changes: 4 additions & 2 deletions providers/cline-pass/models/cline-pass/deepseek-v4.1-flash.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,9 +1,11 @@
# https://api.cline.bot/api/v1/ai/cline/recommended-models (clinePass id cline-pass/deepseek-v4.1-flash, accessed 2026-09-11)
# https://docs.cline.bot/getting-started/clinepass (ClinePass reference pricing cites DeepSeek API pricing for Flash quota)
# https://api-docs.deepseek.com/quick_start/pricing (DeepSeek-V4.1-Flash off-peak USD/MTok; ClinePass docs table still lists deepseek-v4-flash only)
# Reasoning effort values match other ClinePass DeepSeek entries on this host.
# https://api-docs.deepseek.com/guides/thinking_mode/ (OpenAI: thinking.type + reasoning_effort low|high|max; medium/xhigh map to high)
# Effort: reasoning_effort = none|low|high|max (none disables thinking; matches DeepSeek Flash graded set + off)
# ClinePass is an OpenAI-compatible relay; do not advertise GPT-style medium/xhigh.
base_model = "deepseek/deepseek-v4.1-flash"
reasoning_options = [{ type = "effort", values = ["none", "low", "medium", "high", "xhigh"] }]
reasoning_options = [{ type = "effort", values = ["none", "low", "high", "max"] }]

[cost]
input = 0.15
Expand Down
6 changes: 5 additions & 1 deletion providers/cline-pass/models/cline-pass/glm-5.2.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,9 @@
# https://docs.cline.bot/getting-started/clinepass
# Lab Zhipu GLM-5.2: effective effort high|max; none|minimal skip thinking
# (providers/zhipuai/models/glm-5.2.toml; https://docs.bigmodel.cn/cn/guide/capabilities/thinking)
# Effort: reasoning_effort = none|high|max
base_model = "zhipuai/glm-5.2"
reasoning_options = [{ type = "effort", values = ["none", "low", "medium", "high", "xhigh"] }]
reasoning_options = [{ type = "effort", values = ["none", "high", "max"] }]

[cost]
input = 1.4
Expand Down
5 changes: 4 additions & 1 deletion providers/cline-pass/models/cline-pass/kimi-k2.6.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,8 @@
# https://docs.cline.bot/getting-started/clinepass
# Lab Moonshot K2.6: thinking on/off only (providers/moonshotai/models/kimi-k2.6.toml)
# Toggle: thinking / reasoning enabled|disabled (OpenAI-compatible relay; no graded effort on lab)
base_model = "moonshotai/kimi-k2.6"
reasoning_options = [{ type = "effort", values = ["none", "low", "medium", "high", "xhigh"] }]
reasoning_options = [{ type = "toggle" }]

[cost]
input = 0.95
Expand Down
5 changes: 4 additions & 1 deletion providers/cline-pass/models/cline-pass/kimi-k2.7-code.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,8 @@
# https://docs.cline.bot/getting-started/clinepass
# Lab Moonshot K2.7 Code: always-on reasoning, no caller control
# (providers/moonshotai/models/kimi-k2.7-code.toml; openrouter peer also [])
base_model = "moonshotai/kimi-k2.7-code"
reasoning_options = [{ type = "effort", values = ["none", "low", "medium", "high", "xhigh"] }]
reasoning_options = []

[cost]
input = 0.95
Expand Down
4 changes: 3 additions & 1 deletion providers/cline-pass/models/cline-pass/kimi-k3.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,9 @@
# https://docs.cline.bot/getting-started/clinepass (accessed 2026-07-22)
# Model ID cline-pass/kimi-k3; reference pricing $3 / $15 / $0.30 per 1M tokens (input/output/cache read).
# Lab Moonshot K3: thinking.type + output_config.effort low|high|max (providers/moonshotai/models/kimi-k3.toml)
# Effort: reasoning_effort = none|low|high|max (none disables; no medium/xhigh)
base_model = "moonshotai/kimi-k3"
reasoning_options = [{ type = "effort", values = ["none", "low", "medium", "high", "xhigh"] }]
reasoning_options = [{ type = "effort", values = ["none", "low", "high", "max"] }]

[cost]
input = 3
Expand Down
5 changes: 4 additions & 1 deletion providers/cline-pass/models/cline-pass/mimo-v2.5-pro.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,8 @@
# https://docs.cline.bot/getting-started/clinepass
# Lab Xiaomi MiMo-V2.5-Pro: thinking on/off only (providers/xiaomi/models/mimo-v2.5-pro.toml)
# Toggle: thinking / reasoning enabled|disabled (no graded effort on lab)
base_model = "xiaomi/mimo-v2.5-pro"
reasoning_options = [{ type = "effort", values = ["none", "low", "medium", "high", "xhigh"] }]
reasoning_options = [{ type = "toggle" }]

[cost]
input = 1.74
Expand Down
5 changes: 4 additions & 1 deletion providers/cline-pass/models/cline-pass/mimo-v2.5.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,8 @@
# https://docs.cline.bot/getting-started/clinepass
# Lab Xiaomi MiMo-V2.5: thinking on/off only (providers/xiaomi/models/mimo-v2.5.toml)
# Toggle: thinking / reasoning enabled|disabled (no graded effort on lab)
base_model = "xiaomi/mimo-v2.5"
reasoning_options = [{ type = "effort", values = ["none", "low", "medium", "high", "xhigh"] }]
reasoning_options = [{ type = "toggle" }]

[cost]
input = 0.14
Expand Down
5 changes: 4 additions & 1 deletion providers/cline-pass/models/cline-pass/minimax-m3.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,8 @@
# https://docs.cline.bot/getting-started/clinepass
# Lab MiniMax-M3: thinking on/off only (providers/minimax/models/MiniMax-M3.toml)
# Toggle: thinking / reasoning enabled|disabled (no graded effort on lab)
base_model = "minimax/MiniMax-M3"
reasoning_options = [{ type = "effort", values = ["none", "low", "medium", "high", "xhigh"] }]
reasoning_options = [{ type = "toggle" }]

[cost]
input = 0.3
Expand Down
6 changes: 5 additions & 1 deletion providers/cline-pass/models/cline-pass/qwen3.7-max.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,9 @@
# https://docs.cline.bot/getting-started/clinepass
# Lab Alibaba Qwen3.7 Max: enable_thinking + thinking_budget (providers/alibaba/models/qwen3.7-max.toml)
# ClinePass OpenAI-compatible Chat Completions path: peers (OpenRouter) expose on/off only, not budget_tokens.
# Toggle: enable_thinking / reasoning enabled|disabled
base_model = "alibaba/qwen3.7-max"
reasoning_options = [{ type = "effort", values = ["none", "low", "medium", "high", "xhigh"] }]
reasoning_options = [{ type = "toggle" }]

[cost]
input = 2.5
Expand Down
6 changes: 5 additions & 1 deletion providers/cline-pass/models/cline-pass/qwen3.7-plus.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,9 @@
# https://docs.cline.bot/getting-started/clinepass
# Lab Alibaba Qwen3.7 Plus: enable_thinking + thinking_budget (providers/alibaba/models/qwen3.7-plus.toml)
# ClinePass OpenAI-compatible Chat Completions path: peers (OpenRouter) expose on/off only, not budget_tokens.
# Toggle: enable_thinking / reasoning enabled|disabled
base_model = "alibaba/qwen3.7-plus"
reasoning_options = [{ type = "effort", values = ["none", "low", "medium", "high", "xhigh"] }]
reasoning_options = [{ type = "toggle" }]

[cost]
input = 0.4
Expand Down
8 changes: 4 additions & 4 deletions providers/cline-pass/models/cline-pass/qwen3.8-max.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,11 +1,11 @@
# https://docs.cline.bot/getting-started/clinepass (accessed 2026-08-24)
# Model ID cline-pass/qwen3.8-max; reference pricing $2 / $6 / $0.25 per 1M tokens (input/output/cache read), cache write $2.5.
# Lab Alibaba Qwen3.8 Max: enable_thinking + reasoning_effort low|medium|xhigh
# (providers/alibaba/models/qwen3.8-max.toml; https://docs.qwencloud.com/developer-guides/text-generation/thinking)
# Effort includes none for off (hybrid model); no high (lab maps high→xhigh; native graded set is low|medium|xhigh)
base_model = "alibaba/qwen3.8-max"
structured_output = true

[[reasoning_options]]
type = "effort"
values = ["minimal", "low", "medium", "high", "xhigh"]
reasoning_options = [{ type = "effort", values = ["none", "low", "medium", "xhigh"] }]

[cost]
input = 2
Expand Down