can1357 6e0b9d34f2 feat(catalog): increased maxTokens to 65,536 for Kimi K2.7-Code models on Fireworks ...

- Updated models.json to set maxTokens to 65536 for Kimi K2.7-Code entries.
- Adjusted openai-compat logic to use a new 65,536 ceiling for K2.7-Code while preserving the 32,768 cap for older K2 models.
- Modified tests to verify the new token limit and ensure K2.7-Code models are excluded from the older cap.

2026-07-17 03:49:38 +02:00

..

compat

merged PR #5408: fix(catalog): widen GLM coding-plan idle timeout to OpenCode gateways

2026-07-16 03:32:04 +02:00

discovery

Merge PR #4975: fix(cursor): preserve max-mode flag (@roboomp)

2026-07-14 18:31:32 +02:00

identity

fix(ai): omitted unsupported reasoning sampling params

2026-07-15 19:30:53 +00:00

provider-models

feat(catalog): increased maxTokens to 65,536 for Kimi K2.7-Code models on Fireworks

2026-07-17 03:49:38 +02:00

wire

Merge remote-tracking branch 'origin/farm/acd088b5/github-copilot-business-vision'

2026-07-11 07:40:14 +02:00

build.ts

feat(catalog): updated model configurations and reasoning tiers

2026-06-22 05:18:47 +02:00

effort.ts

feat: introduced max reasoning effort tier across model and rpc systems

2026-07-10 13:39:42 +02:00

fireworks-model-id.ts

feat(catalog): supported fireworks fast serving path

2026-06-20 09:21:01 +02:00

hosts.ts

feat(catalog): integrated baseten provider and updated model definitions

2026-07-03 06:04:31 +02:00

index.ts

feat(ai): add CoreWeave Serverless Inference provider

2026-06-26 09:13:41 -04:00

model-cache.ts

fix(catalog): stopped promoting legacy v2 model cache rows

2026-07-01 13:56:48 +00:00

model-manager.ts

Merge PR #4959: fix(catalog): preserve static limits on cache fingerprint mismatch (@roboomp)

2026-07-14 18:34:13 +02:00

model-thinking.ts

fix(catalog): classify Z.ai GLM-5.2 anthropic-messages endpoint as budget-effort with high/max

2026-07-10 14:43:31 -07:00

models.json

feat(catalog): increased maxTokens to 65,536 for Kimi K2.7-Code models on Fireworks

2026-07-17 03:49:38 +02:00

models.json.d.ts

refactor(catalog)!: split model catalog from pi-ai

2026-06-10 04:06:57 +02:00

models.ts

fix(ai): separated codex orchestration usage

2026-07-03 16:44:12 +00:00

types.ts

merged PR #5608: fix(ai): omit sampling params for OpenAI reasoning models

2026-07-16 03:31:57 +02:00

utils.ts

fix(catalog): unified ssl fetch overrides under discoveryFetch

2026-07-02 02:40:07 +02:00

variant-collapse.ts

feat: introduced max reasoning effort tier across model and rpc systems

2026-07-10 13:39:42 +02:00