Skip to content

Resolve dev conflicts in #5563 - #7908

Closed
superhighfives wants to merge 2033 commits into
anomalyco:fix/workers-reasoning-sync-aidenfrom
superhighfives:fix/workers-reasoning-sync-aiden
Closed

Resolve dev conflicts in #5563#7908
superhighfives wants to merge 2033 commits into
anomalyco:fix/workers-reasoning-sync-aidenfrom
superhighfives:fix/workers-reasoning-sync-aiden

Conversation

@superhighfives

@superhighfives superhighfives commented Sep 23, 2026

Copy link
Copy Markdown
Contributor

Fixes the merge conflicts blocking #5563.

Agent notes

Merged current dev and reconciled overlapping Workers AI reasoning and header logic. Model-specific Search metadata takes precedence, per-model schemas provide fallback controls, and generated comments preserve authored rationale.

Verified with the focused Workers AI tests (27 passed) and bun validate. The broader core suite has five unrelated failures on current dev (394 passed).

opencode-agent Bot and others added 30 commits September 21, 2026 14:29
@opencode-agent
chore(sync): update OpenRouter model catalog (anomalyco#7634)
e379bc9 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update Kilo model catalog (anomalyco#7633)
074421e 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@0genlab chenxue @claude
fix(deepseek): align V4 Pro effort with GA documentation (anomalyco#7564
88bb3d9 
)

* fix(deepseek): align V4 Pro effort with GA documentation

* chore: retrigger review workflow

The previous review run failed while installing opencode
("Failed to fetch version information"), so no code review ran.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

---------

Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
@pc-gs
fix(zai): add missing glm-4.6v-flash provider model (anomalyco#7592)
843685c 
Resolves broken zhipuai symlink to zai/glm-4.6v-flash.toml.

Fixes anomalyco#1470

Co-authored-by: Pedro Santos <18473317+pc-gs@users.noreply.github.com>
@0genlab chenxue
feat(aihubmix): add step-3.7-flash (anomalyco#7559)
d0b5222 
Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
@0genlab chenxue
feat(aihubmix): add minimax-m3 (anomalyco#7558)
137714d 
Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
@0genlab chenxue
feat(aihubmix): add gemini-3.1-flash-image (anomalyco#7556)
1da5b84 
Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
@0genlab chenxue
feat(aihubmix): add hy3 (anomalyco#7555)
1fc9f0f 
Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
@0genlab chenxue
feat(aihubmix): add muse-spark-1.1 (anomalyco#7554)
4c5584c 
* feat(aihubmix): add muse-spark-1.1

* docs(aihubmix): cite Muse Spark 1.1 audio input declaration

---------

Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
@0genlab chenxue
feat(aihubmix): add gemini-3.5-flash-lite (anomalyco#7553)
d5fb44d 
* feat(aihubmix): add gemini-3.5-flash-lite

* docs(aihubmix): cite exact catalog price for Gemini 3.5 Flash Lite

---------

Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
@0genlab chenxue
feat(aihubmix): add gemini-3.1-flash-lite-image (anomalyco#7552)
5d137a9 
Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
@0genlab chenxue
feat(aihubmix): add muse-spark-1.2 (anomalyco#7551)
90db5bf 
Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
@0genlab chenxue
feat(aihubmix): add qwen3.8-omni-flash (anomalyco#7550)
21e40fe 
Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
@Pessimist228
llmtech: rename model to nvidia/Qwen3.8-27B-NVFP4 (anomalyco#7549)
0578dc2 
* llmtech: canonical model id is nvidia/Qwen3.8-27B-NVFP4

The served build is NVIDIA's own NVFP4 checkpoint, not the community one.
The previous id stays accepted by the API as a synonym, so nothing breaks
for anyone who has not switched yet.

* llmtech: drop the old unsloth/ model path
@fanweixiao
feat(vivgrid): add viv-fast and deepseek-v4.1-flash models (anomalyco…
ba1ee71 
…#7626)

- viv-fast: new Vivgrid first-party SLM (1M context, 256K output,
  text+image input, reasoning effort low/high/max), lab entry under
  models/vivgrid/ plus provider pricing at $0.13/$0.40 with $0.05
  cache read
- deepseek-v4.1-flash: base_model override on deepseek/deepseek-v4.1-flash
  with Vivgrid pricing ($0.31/$1.23, $0.01 cache read), DeepSeek-native
  reasoning controls (thinking toggle + effort low/high/max) and
  reasoning_content interleaved, matching the existing deepseek-v4-flash
  entry on the same surface
@BUKOWSKIREAL
feat(stepfun-step-plan): add step-5-preview (anomalyco#7544)
5d395d4 
StepFun's new flagship base model, available on the Step Plan gateway
(api.stepfun.com/step_plan/v1). Lab metadata: 1M-token context/input/
output, text/image/video input, reasoning + tool calling + structured
output, closed weights. Provider override: reasoning_effort accepts
low/medium/high; reasoning side channel is reasoning_content.

Sources:
- https://platform.stepfun.com/docs/zh/guides/models/step-5-preview
- https://api.stepfun.com/step_plan/v1/models (accessed 2026-09-20)

bun validate passes.

Co-authored-by: BUKOWSKIREAL <BUKOWSKIREAL@users.noreply.github.com>
@opencode-agent
chore(sync): update Kilo model catalog (anomalyco#7637)
9b23514 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update OpenRouter model catalog (anomalyco#7635)
c49c7b1 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@0genlab chenxue
feat(aihubmix): add longcat-2.0 (anomalyco#7557)
1e4a5a8 
Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
@noahchen2002
feat(siliconflow): sync model catalog with SiliconFlow international …
12465a3 
…site (anomalyco#7625)

* feat(siliconflow): sync catalog with SiliconFlow international site (2026-09)

Add 13 models: DeepSeek-V4-Flash-Vision-Exp, DeepSeek-V4-Pro-0813,
DeepSeek-V4-Flash-0731, Qwen3.8-2.4T-A95B, GLM-5.3, GLM-5.3-Flash,
Kimi-K3, Kimi-K2.7-Code, MiniMax-M3, gemma-4-12B-it, Nex-N2-Pro,
meituan-longcat/LongCat-2.0, tencent/Hy3 (replaces Hy3-preview).

Remove 5 models no longer listed on SiliconFlow international:
ERNIE-4.5-300B-A47B, Qwen3-VL-235B-A22B-Instruct/Thinking,
Qwen3-235B-A22B-Thinking-2507, Hy3-preview.

Update pricing to current rates for DeepSeek-V3.2, DeepSeek-V4-Flash,
DeepSeek-V4-Pro, Kimi-K2.6, GLM-5.1, GLM-5.2.

Add lab metadata for google/gemma-4-12b-it and nex-agi/nex-n2-pro.

Sources: https://www.siliconflow.com/models,
https://www.siliconflow.com/pricing,
https://docs.siliconflow.com/en/release-notes/overview.md

* fix(siliconflow): address auto-review items on reasoning options

- LongCat-2.0: drop toggle (SF does not enumerate enable_thinking for
  LongCat), keep budget_tokens with leading wire-path comment
- MiniMax-M3, Qwen3.8-2.4T-A95B: add leading wire-path comments citing
  SF's thinking_budget surface; SF exposes no reasoning_effort param
- Hy3: drop stale open_weights=false override inherited from Hy3-preview
  (lab entry tencent/hy3 is open weights)
- nex-agi/nex-n2-pro lab: attachment=true, consistent with image input
  modality

* fix(siliconflow): address round-2 auto-review items

- Narrow inherited modalities to SF's actual surface: MiniMax-M3,
  Kimi-K3, Kimi-K2.7-Code, GLM-5.3-Flash now serve text+image only
  (per each model page's "Support image input: Yes"; no video/pdf on SF)
- Hy3: drop redundant [modalities] (identical to lab), document that
  262K/262K limits are SF-advertised deltas vs lab 256K/128K
- GLM-5.1: cite exact price row (input 1.19 / cached 0.6 / output 3.74
  from the GLM-5.1 model page) - cache_read=0.6 is SF's published price
- V4-Pro / K2.7-Code / GLM-5.2: note costs are published USD list
  prices from siliconflow.com/pricing, not FX conversions

* fix(siliconflow): address round-3 auto-review items

- GLM-5.3 / GLM-5.3-Flash: effort low|high|max, matching the first-party
  zhipuai GLM-5.3 entry (5.2's high|max was generation-specific)
- gemma-4-12B-it: drop [limit].context (identical to lab), keep only the
  real output delta

---------

Co-authored-by: NoahChen <288864333+noahchen2002@users.noreply.github.com>
@opencode-agent
chore(sync): update Eden AI model catalog (anomalyco#7636)
e64a289 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@knorby
Update neuralwatt catalog: add 8 models, deprecate retired GLM-5.2 fa…
66635ce 
…mily and V4 Pro (anomalyco#7482)
@opencode-agent @rekram1-node
chore(sync): update Hugging Face model catalog (anomalyco#7562)
6c97e53 
* chore(sync): update Hugging Face model catalog

* fix(huggingface): add Hy4 reasoning options

---------

Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
Co-authored-by: rekram1-node <rekram1-node@users.noreply.github.com>
@github-actions
fix: [missing-model] xai: grok-imagine-image-quality (anomalyco#7391)
7c89bba 
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
@0genlab chenxue @claude
feat(aihubmix): add gpt-6-astra (anomalyco#7422)
3afe1fe 
AIHubMix serves this route but the catalog carried no aihubmix entry for
it. Cost comes from the provider's public model listing and the reasoning
controls from its published model-data index; limit, modalities,
attachment and tool_call inherit from models/openai/gpt-6-astra.toml rather than being
repeated here.

Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
@0genlab chenxue @claude
feat(aihubmix): add gemini-3.8-flash (anomalyco#7423)
7815e59 
* feat(aihubmix): add gemini-3.8-flash

AIHubMix serves this route but the catalog carried no aihubmix entry for
it. Cost comes from the provider's public model listing and the reasoning
controls from its published model-data index; limit, modalities,
attachment and tool_call inherit from models/google/gemini-3.8-flash.toml rather than being
repeated here.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* aihubmix/gemini-3.8-flash: drop budget_tokens from reasoning_options

The bare budget_tokens entry came from a generic multi-protocol field
dump, not from model-level evidence: first-party
providers/google/models/gemini-3.8-flash.toml and the AIHubMix peer
gemini-3.7-flash are both effort-only, and Google's documented
reasoning-token budgets apply to the 2.5 family, not 3.8.

Effort low|medium|high is unchanged and is the model-level control.
The leading comment loses the budget field paths along with it.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

---------

Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
@0genlab chenxue @claude
feat(aihubmix): add hy4-preview (anomalyco#7424)
cd17c40 
AIHubMix serves this route but the catalog carried no aihubmix entry for
it. Cost comes from the provider's public model listing and the reasoning
controls from its published model-data index; limit, modalities,
attachment and tool_call inherit from models/tencent/hy4-preview.toml rather than being
repeated here.

Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
@0genlab chenxue @claude
feat(aihubmix): add muse-spark-1.3 (anomalyco#7425)
8824596 
* feat(aihubmix): add muse-spark-1.3

AIHubMix serves this route but the catalog carried no aihubmix entry for
it. Cost comes from the provider's public model listing and the reasoning
controls from its published model-data index; limit, modalities,
attachment and tool_call inherit from models/meta/muse-spark-1.3.toml rather than being
repeated here.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* aihubmix/muse-spark-1.3: drop budget_tokens from reasoning_options

The bare budget_tokens entry came from a generic multi-protocol field
dump rather than a model-level control: first-party Meta and the
established relays of meta/muse-spark-1.3 are effort-only, and the
same-host AIHubMix effort-only peers omit it as well.

The effort set stays at minimal|low|medium|high|xhigh. `max` is
deliberately not listed: Meta's 1.3 launch post says "Previously
available reasoning modes are available today with max reasoning
coming shortly after we finish additional safety testing", so the
"Muse Spark 1.3 (max)" benchmark label is not an API availability
statement yet.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

---------

Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
@0genlab chenxue @claude
feat(aihubmix): add deepseek-v4-flash-0731-fast (anomalyco#7429)
96655bf 
AIHubMix serves this route but the catalog carried no aihubmix entry for
it. Cost comes from the provider's public model listing and the reasoning
controls from its published model-data index; limit, modalities,
attachment and tool_call inherit from models/deepseek/deepseek-v4-flash-0731.toml rather than being
repeated here.

Co-authored-by: chenxue <DDU1222@users.noreply.github.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
@KastanDay
fix(cloudflare): correct catalogue reasoning efforts & base_models (a…
63cca5f 
…nomalyco#7454)

* fix(cloudflare): refresh catalogue reasoning and canonical links

* fix(cloudflare): sync GLM 4.7 toggle and expand reasoning tables

* fix(cloudflare): cap oversized context and output limits

* fix(cloudflare): inherit SEA-LION description and clarify reasoning sources

* docs(cloudflare): cite GLM thinking control at its catalogue entry

* docs(cloudflare): cite dated DeepSeek effort controls

* fix(cloudflare): omit fixed Kimi K2.7 effort control
opencode-agent Bot and others added 26 commits September 23, 2026 09:30
@opencode-agent
chore(sync): update Kilo model catalog (anomalyco#7860)
1eb046d 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update OpenRouter model catalog (anomalyco#7866)
7b48252 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update Kilo model catalog (anomalyco#7865)
e50d389 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update NanoGPT model catalog (anomalyco#7867)
31a4a59 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update OpenRouter model catalog (anomalyco#7870)
fb4e56d 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update Kilo model catalog (anomalyco#7869)
137601b 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update Kilo model catalog (anomalyco#7880)
3beb7f4 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update OpenRouter model catalog (anomalyco#7881)
c2cb379 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@MrMushrooooom
feat(opencode): add Space Bunny Free
6280e4b
@opencode-agent
chore(sync): update Kilo model catalog (anomalyco#7884)
8cf8b71 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update OpenRouter model catalog (anomalyco#7885)
f42d1b1 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update Eden AI model catalog (anomalyco#7887)
15a35f2 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update Vercel AI Gateway model catalog (anomalyco#7886)
4e7a144 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update NanoGPT model catalog (anomalyco#7888)
ad2ee37 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update Eden AI model catalog (anomalyco#7892)
1edbb1d 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update OpenRouter model catalog (anomalyco#7889)
b62670a 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update DevPass (LLM Gateway) model catalog (anomalyco#7891)
f203644 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update LLM Gateway model catalog (anomalyco#7890)
59f2e56 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update NanoGPT model catalog (anomalyco#7896)
d1f1da9 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update Kilo model catalog (anomalyco#7897)
a0d3c7c 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update OpenRouter model catalog (anomalyco#7895)
806fdff 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update Venice model catalog (anomalyco#7894)
f5f8f45 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update Kilo model catalog (anomalyco#7898)
963b85b 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update OpenRouter model catalog (anomalyco#7899)
f2fee37 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@opencode-agent
chore(sync): update Vercel AI Gateway model catalog (anomalyco#7900)
bf47e10 
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
@superhighfives
Merge remote-tracking branch 'origin/dev' into fix/workers-reasoning-…
1601c72 
…sync-aiden

# Conflicts:
#	packages/core/src/sync/index.ts
#	packages/core/src/sync/providers/cloudflare-workers-ai.ts
@superhighfives

superhighfives commented Sep 23, 2026

Copy link
Copy Markdown
Contributor Author

Closing because this targeted the stale PR branch and included the intervening dev history. Replacing it with a clean PR based on current dev.

@superhighfives
superhighfives deleted the fix/workers-reasoning-sync-aiden branch September 23, 2026 21:19
@superhighfives superhighfives mentioned this pull request Sep 23, 2026
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.