Ling-3.0-flash: benchmark results
InclusionAI's open 124B MoE (5.1B active) with native hybrid-linear attention and hybrid reasoning, built for high-throughput agent workloads (July 2026). Provider: InclusionAI. Released 2026-07-23. Access: Open.
Aggregate Rating 1696 ± 12, rank #202 of 1640 rated models, from 220 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| ArtifactsBenchmark | 77 | Average Score | 100 |
| LLM Stats (Multi-IF) | 87.7 | Score (%) | 100 |
| Multi-IF | 87.7 | Average (self-reported) | 96.4 |
| Nejumi 4 - MT-Bench (Japanese) - Coding | 100 | Judge rating (1-10, x10) | 91.5 |
| Nejumi 4 - MT-Bench (Japanese) - STEM | 100 | Judge rating (1-10, x10) | 90.4 |
| Nejumi 4 - BFCL - Multi-Turn | 56.67 | Accuracy (%) | 88.7 |
| Nejumi 4 - jaster (0-shot) - JSQuAD | 81.12 | Character F1 (x100) | 87.2 |
| Nejumi 4 - jaster (0-shot) - JaMP | 77 | Exact match (%) | 86.2 |
| AA-Omniscience Hallucination Rate | 44.05 | Hallucination Rate (%) | 84.8 |
| Nejumi 4 - jaster (2-shot) - JaNLI | 100 | Exact match (%) | 84.8 |
| Nejumi 4 - jaster (0-shot) - JaNLI | 100 | Exact match (%) | 82.6 |
| Tinybird AI SQL Benchmark - First-Attempt Success Rate | 98 | Questions answered with a valid query on the first attempt ( | 80.4 |
Interactive version: theaggregate.ai/model?slug=ling-3-0-flash · How It Works · Data refreshed daily, snapshot 2026-10-13.