pos-ei-don 879c686ab1 feat(sglang): map reasoning_effort to per-level thinking budgets ...

The Go side already forwards the request's effort as
Metadata["reasoning_effort"], but the sglang backend only read
enable_thinking: every level got the same model-level thinking_budget,
so low/medium/high produced identical reasoning.

- New model option `effort_budgets:low=…;medium=…;high=…` picks the
  budget of the request's level; unlisted levels and requests without an
  effort keep `thinking_budget`. Parsing reuses _parse_thinking_budget.
- A level listed in effort_budgets turns reasoning on for that request,
  so levels the Go side does not know cannot silently disable it.
- ApplyReasoningEffort now treats `xhigh` and `max` as thinking levels.
  Before, they left DisableReasoning unset, so with a model default of
  "off" asking for more reasoning produced none.

Signed-off-by: pos-ei-don <1822533+pos-ei-don@users.noreply.github.com>

2026-10-14 03:07:52 +02:00

..

_index.en.md

docs: fix dead anchors and a dead section link (#11885)

2026-09-05 23:47:13 +02:00

advanced-usage.md

docs: point the config example link at the configurations directory

2026-09-27 19:06:33 +00:00

embedded-routing-log.md

feat(router): route with native decision models (#12449)

2026-10-04 09:34:21 +02:00

model-configuration.md

feat(sglang): map reasoning_effort to per-level thinking budgets

2026-10-14 03:07:52 +02:00

private-sources.md

feat(gallery): galleries published as OCI artifacts (#12167)

2026-09-21 09:23:05 +02:00

reverse-proxy-tls.md

docs: add Apache APISIX reverse proxy example (#11294)

2026-08-28 08:51:38 +02:00

vram-management.md

fix(distributed): exclude pinned models from cluster eviction and idle scale-down

2026-09-11 23:08:26 +00:00