# Qwen3.8 Flash (Alibaba)

LLM, Alibaba, no reasoning. Verbalized probabilities through the Jevals adapter, via OpenRouter · Alibaba.

| Board | Decision Score (95% range) | Rank | Accuracy | Calibration gap (ECE, pts) | Hand-off at 95% | $ per 1k decisions | Speed p95 |
| --- | --- | --- | --- | --- | --- | --- | --- |
| [noul (PubMedQA)](https://jevals.com/noul/) | 62.4 (50.9 to 72.5) | rank 2 of 8 | 89.7% | 2.5 | — | $0.086 | 3.5 s |
| [choice (Banking77)](https://jevals.com/choice/) | 62.4 (55.9 to 68.9) | tied for rank 4 of 8 | 76.8% | 9.5 | — | $0.12 | 3.1 s |
| [score (HelpSteer2 helpfulness)](https://jevals.com/score/) | −1.4 (−15.4 to 11.5) | tied for rank 3 of 8 | 36.1% | 25.8 | — | $0.13 | 2.6 s |

Decision Score: 100 = perfect, 0 = guessing the label base rates, below 0 = worse than that. Ranks: 1 + the number of rows significantly better (paired item bootstrap, 95%); rows that cannot be told apart share a rank. Hand-off at 95%: the share of decisions a model can take alone, at its own confidence threshold, while staying at least 95% right.

Setup: model id qwen/qwen3.8-flash; served as qwen/qwen3.8-flash; adapter jevals-llm 0383a0e3e592; reasoning none; temperature provider default; price as of 2026-09-18.

Compare: [Jev vs Qwen3.8 Flash](https://jevals.com/compare/jev-vs-qwen3.8-flash/)

Source: https://jevals.com/models/qwen3.8-flash/ · release 2026-09-18, suite 0.1.0 · data (CC-BY-4.0): https://jevals.com/data/releases/2026-09-18/board.json · cite as: Jevals (jevals.com), release 2026-09-18, suite 0.1.0. CC-BY-4.0.