Ant's InclusionAI Ships Ling-3.1-flash With 560B Parameters
TL;DR
- All 8 publicly visible benchmark scores for Ling-3.1-Flash are vendor-reported from Ant's own launch screenshots; no third-party evaluation has occurred.
- Vercel added Ling-3.1-Flash to AI Gateway on launch day, giving developers API access before Ant published pricing or model weights.
- OrcaRouter documents that Ling-3.0-Flash had a two-week gap between announcement and open-source release, making a similar delay the base-case expectation for Ling-3.1.
Ant Group's InclusionAI has released Ling-3.1-flash, a language model with 560 billion total parameters and about 25 billion activated for each token. Per TechNode, the company is "positioning it for agent tasks, search, office software and specialist applications."
The full design targets a context window of up to one million tokens, but the two-week free trial launched alongside the model is capped at 256,000. InclusionAI says it "plans to enable the larger window and release the model as open source after the trial."
The report names no spokesperson and publishes no benchmark numbers. It fits a steady run of Chinese-lab model drops we have been tracking this quarter.
What others are reporting
-
Vercel Read →
Vercel integrated the model same-day with step-by-step setup for Claude Code, Codex, and fx, clarifying the free trial model ID structure to prevent unexpected billing.
The model is designed for coding, multi-step analysis, and agents that use tools, including workflows involving long documents, code, and extended task histories.
-
OrcaRouter Read →
Catalogs the gap between announcement and verifiable product: no weights, no external evals, and a historical parallel to Ling-3.0-Flash's two-week release lag.
Ling-3.1-flash is not an open release. It is an announced one. The two are not close, and the gap between them is exactly where a reader can get burned.
-
ThreatFrontier Read →
Comparative architecture and licensing analysis against two competing Chinese flash MoE models, with enterprise data-residency and compliance framing.
InclusionAI has not published a single benchmark for Ling 3.1 Flash, so any head-to-head ranking of it today is guesswork.
-
BenchLM Read →
Independent benchmark tracker lists 8 vendor-reported scores from Ant's launch screenshots but assigns no overall rank, flagging the absence of third-party evaluation.
Ling 3.1 Flash by InclusionAI has 8 source-displayable benchmark rows but no public overall score. It remains unranked.
Originally reported by technode.com
Read the original article →Original headline: Ant Group's InclusionAI Ships Ling-3.1-Flash, a 560B MoE with 25B Active and 1M-Token Context