Intel / post

Ling 3.1 Flash doubles its Intelligence Index score to 41

Source: x.com
Date: Published Oct 6, 2026

Why it matters

Ling 3.1 Flash (560B total, 25B active, 1M ) jumps from 20 to 41 on the Intelligence Index, with gains on agentic and terminal tasks. Weights are expected soon, which makes it a candidate open model for agent workloads.

Key takeaways · AI-distilled

Artificial Analysis reports Ling 3.1 Flash scores +2 on AA-Omniscience, up from -18: accuracy rose from 18% to 29% and fell from 44% to 38% at a similar attempt rate, so the gain comes from knowing more, not abstaining more.
On agentic tests it posts a GDPval-AA v2 Elo of 1,622 and an AA-Briefcase Elo of 1,400, plus 62% on AutomationBench-AA and 33% on Terminal-Bench v4.0, which Artificial Analysis says rivals GLM-5.3-Flash and Gemini 3.8 Flash.
It is far larger than Ling 3.0 Flash (124B total, 5.1B active) and pricier at $0.30/$0.90 per 1M input/output tokens, yet used 16% fewer output tokens (218M vs 261M) to run the Intelligence Index.
At $0.99 per task it costs well above GLM-5.3-Flash ($0.42) and DeepSeek V4.1 Flash ($0.32) but below Gemini 3.8 Flash (High) at $1.24. It is available through Novita AI, with weights coming soon.

Terms in this piece · Glossary

reasoning model — A model trained to think — generating extended internal reasoning before answering — trading time and tokens for accuracy on hard problems.
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
hallucination — When a model states something false with full confidence — inventing facts, citations, or APIs that don't exist.
open weights — A model whose trained parameters are published for anyone to download and run — unlike API-only models you can access but never possess.

Read the source x.com