run_2026-08-27_nightly
Nightly · v3.3 · 2026-08-27
Claude Sonnet 5.5 · bge-m3 · hybrid (BM25 + dense, RRF) · support-prompt v6 on support-golden-v4 v4.6
Stratified slice of support-golden-v4 (430 items).
Pass rate
81.9%
Faithfulness
0.86
Answer correctness
0.84
Context precision
0.73
Context recall
0.80
Latency p95
3.8s
Failures by category
78 failed samples · largest share: unsupported
| Category | Share | vs reference |
|---|---|---|
| Unsupported claims | 25.6% | −0.3 pts |
| Incorrect factual answers | 20.5% | +0.8 pts |
| Missing context | 21.8% | +0.8 pts |
| Retrieval mismatch | 20.5% | −0.5 pts |
| Instruction-following failures | 9.0% | +0.3 pts |
| Other | 2.6% | −1.1 pts |
Run details
- Dataset
- support-golden-v4 v4.6 · 430 items
- Generator
- Claude Sonnet 5.5
- Retriever
- bge-m3 · hybrid (BM25 + dense, RRF)
- Prompt
- support-prompt v6
- Judge
- LLM judge · Claude Opus 5.5 · rubric v3
- Started
- Aug 27, 02:00 UTC
- Duration
- 44m 6s
- Triggered by
- Scheduler
- Tokens / request
- 3,980 in · 221 out
- Compared with
- Nightly · v3.3 · 2026-08-26
- nightly
- golden-slice
All metrics
Against Nightly · v3.3 · 2026-08-26
| Metric | Group | This run | Reference | Change |
|---|---|---|---|---|
| Pass rateShare of samples the judge marked as passing all rubric checks. | Outcome | 81.9% | 80.3% | +1.6 pts |
| FaithfulnessShare of answer claims supported by the retrieved context. | Generation | 0.86 | 0.87 | −0.01 |
| Answer correctnessAgreement with the gold answer on facts, not style. | Generation | 0.84 | 0.85 | −0.01 |
| Instruction adherenceRespect for tone, format, length, language and policy constraints. | Generation | 0.89 | 0.90 | −0.01 |
| Context precisionRank-weighted precision of relevant chunks passed to the generator. | Retrieval | 0.73 | 0.74 | −0.01 |
| Context recallShare of gold-answer statements supported by the retrieved context. | Retrieval | 0.80 | 0.79 | +0.01 |
| Recall@5Share of gold passages found in the top 5 retrieved chunks. | Retrieval | 0.83 | 0.82 | +0.01 |
| MRRMean reciprocal rank of the first relevant chunk. | Retrieval | 0.68 | 0.69 | −0.01 |
| NDCG@10Graded ranking quality of the top 10, normalized by the ideal order. | Retrieval | 0.75 | 0.76 | −0.01 |
| Latency p50Median end-to-end time to final token. | Operational | 1.8s | 1.8s | 0.0s |
| Latency p9595th percentile end-to-end time to final token. | Operational | 3.8s | 3.5s | +0.3s |
| Cost per requestModel, embedding and reranking cost per answered request (USD). | Operational | $0.0046 | $0.0046 | 0.0000 |
Nightly trend
Nightly runs on the support-golden-v4 slice, Jul 11 to Oct 8. This run is one point in the series.
Across datasets
support-assistant v3.3 on every dataset in the project. Rows marked “matrix” come from the comparison matrix rather than a dedicated run.
| Dataset | Items | Pass rate | Faith. | Ctx recall | Recall@5 | p95 | Source |
|---|---|---|---|---|---|---|---|
| support-golden-v4this run | 1,240 | 81.9% | 0.86 | 0.80 | 0.83 | 3.8s | This run |
| billing-edge-cases | 312 | 76.0% | 0.83 | 0.77 | 0.81 | 3.9s | Run · Aug 20 |
| roaming-and-travel | 188 | 75.5% | 0.84 | 0.74 | 0.78 | 3.8s | matrix |
| device-troubleshooting | 426 | 78.9% | 0.85 | 0.78 | 0.82 | 4.4s | matrix |
| account-security | 154 | 82.5% | 0.88 | 0.80 | 0.84 | 3.6s | matrix |
| locale-formatting | 96 | 75.0% | 0.82 | 0.75 | 0.79 | 4.0s | matrix |
| synthetic-hard-negatives | 540 | 70.7% | 0.82 | 0.68 | 0.72 | 3.9s | matrix |
Traced samples
17 samples with full traces (4 failing, 13 passing), stratified by domain. Open a row to inspect retrieval and the judge's reasoning.
My bill says 41,20 €. Is that forty-one euros or four thousand? Billing & payments Passjudge 0.91 |
Can I return a phone I bought online if I opened the box? Returns & warranty Passjudge 0.95 |
Is SIM swap protection turned on by default? Account & security Passjudge 0.88 |
My eSIM QR code says it has already been used. What should I do? eSIM & activation Passjudge 0.89 |
What is the excess on a theft insurance claim? Returns & warranty Passjudge 0.94 |
Is there a restocking fee if I return an opened phone? Returns & warranty FailMissingjudge 0.13 |
How much data can I use in Germany on Unlimited Plus before I get charged? Roaming & travel FailMismatchjudge 0.18 |
My eSIM stopped working after I reset my phone. Do I need to pay for a new one? eSIM & activation FailUnsupportedjudge 0.12 |
Can I keep my old number if I move from prepaid to a contract plan? Plans & upgrades FailIncorrectjudge 0.48 |
Does the Fall 2026 upgrade offer work with a family plan discount? Promotions Passjudge 0.91 |