@miller-tech/uap 2.19.5 → 2.21.0

package/config/model-profiles/qwen38.json

@@ -1,9 +1,9 @@
 {
   "_profile": "qwen38",
   "_description": "Qwen3.8 Flash-Next — local Strata serve-layer deployment for agentic tool-calling workloads. Successor to qwen36 and to the 2026-09-20 GSQ-RCO llama.cpp deployment; the default local executor in UAP routing. Routes through the anthropic-proxy for the tool/finalize guardrails, NOT the inference server directly.",
-  "_engine": "strata serve layer 0.1.37 (python server + native engine binary) on :8080, no systemd unit — launched by /home/cogtek/dev/strata/run-iq3_xxs.sh. Serves llama.cpp-shaped /props, /slots and /v1, plus a JSON /metrics (engine/live/requests — NOT Prometheus text; `uap inference health` auto-detects it). Weights: strata-data pack iq3_xxs (Qwen3.8-Flash-Next-GSQ-RCO-IQ3_XXS, native experts + PLE shard) with MTP speculative drafts (spec 4, spec-min-p 0.5, mtp_max 4) and an auto-sized expert cache (~14.5 GiB, 91% hit rate observed). Model id advertised at /v1/models: qwen3.8-flash-next-iq3_xxs; the server ignores requested model ids entirely, which is why stale client pins (e.g. a resumed session still naming qwen36 ids) keep working.",
+  "_engine": "strata serve layer 0.1.37 (python server + native engine binary) on :8080, no systemd unit — launched by /home/cogtek/dev/strata/run-iq3_xxs.sh. Serves llama.cpp-shaped /props, /slots and /v1, plus a JSON /metrics (engine/live/requests — NOT Prometheus text; `uap inference health` auto-detects it). Weights: strata-data pack iq3_xxs (Swift 1.5 IQ3_XXS since the 2026-10-06 swap; predecessor Qwen3.8-Flash-Next-GSQ-RCO-IQ3_XXS, native experts + PLE shard) with MTP speculative drafts (spec 8, spec-min-p 0.8 since the 2026-10-06 tuning; previously spec 4, spec-min-p 0.5, mtp_max 4) and an auto-sized expert cache. Model id advertised at /v1/models: swift-1.5-iq3_xxs; the server ignores requested model ids entirely, which is why stale client pins (e.g. a resumed session still naming qwen36 ids) keep working.",
   "_engine_history": "SUPERSEDED 2026-10-04: this profile described buun-llama-cpp behind uap-gsq-rco-server.service (2 rails sharing a 229376-cell VBR pool, DFlash2 draft). That unit is now INACTIVE — the box serves the Strata layer on :8080 instead (tuned 2026-10-03: +15% over the llama.cpp fork on agent workloads). The old description was itself a 2026-09-20 correction of a ninfer-serve description. Verify against /metrics before trusting engine claims here; the dead unit's ExecStart still reports -c 229376 to systemctl, and on 2026-10-04 `uap inference health` printed that stale pool size for the live (131072) backend.",
-  "model": "qwen3.8-flash-next-iq3_xxs",
+  "model": "swift-1.5-iq3_xxs",
   "max_tokens": 57344,
   "_max_tokens_comment": "Deliberately generous: 50% of the 114688 window, the same ratio the previous 65536/131072 pairing used. The engine returns chain-of-thought in a SEPARATE reasoning_content field, so reasoning spends completion budget without ever appearing in content. A tight cap therefore does not produce a short answer, it produces an EMPTY one with finish_reason=length — measured previously as ~90% first-turn decode failure that read as bad output format and was really budget truncation. Do not lower this without re-measuring decode compliance. (The proxy additionally clamps tool turns to PROXY_TOOL_TURN_MAX_TOKENS=32768 regardless.)",
   "temperature": 0.3,