Qwen3-Swallow-8B-RL-v0.2 (Thinking): benchmark results

Provider: Alibaba. Access: Open.

Unified ELO 1527 ± 1, rank #816 of 2032 rated models, from 101 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Nejumi 4 - jaster (2-shot) - JSICK89Exact match (%)100
Nejumi 4 - jaster (0-shot) - JSICK84Exact match (%)92.3
Nejumi 4 - jaster (0-shot) - JCoLA (in-domain)79Exact match (%)83.5
Swallow - Japanese MT-Bench - Math98.4Judge Score (normalized, %)82.8
Nejumi 4 - jaster (0-shot) - JaNLI100Exact match (%)82.4
Swallow - Post-trained Japanese - PolyMath High and Top45.6Accuracy (%)73.1
Swallow - English MT-Bench - Extraction80.1Judge Score (normalized, %)72.4
Nejumi 4 - MT-Bench (Japanese) - STEM99.5Judge rating (1-10, x10)71.3
Nejumi 4 - BFCL - Relevance Detection77.78Accuracy (%)70.2
Nejumi 4 - BFCL - Non-Live AST77.78Accuracy (%)64.7
Swallow - Post-trained English - AIME73.3Accuracy (%)64.2
Nejumi 4 - BFCL - Irrelevance Detection85Accuracy (%)61.8

Interactive version: theaggregate.ai/model?slug=qwen3-swallow-8b-rl-v0-2-thinking · How It Works · Data refreshed daily, snapshot 2026-09-26.