David Hendrickson(@TeksEdge):Wow, that's a pretty nice bump for Ling-3.1-flash. Ant introduced the new Ling-3.1-flash-1019 checkpoint, and its Artificial Analysis score jumped Ling-3.1-flash → 41 Ling-3.1-flash-1019 → 44 The highlighted gains come from its updated agentic performance AA-Briefcase ~1400 → 1524 Elo AutomationBench-AA 62% → 68% The existing Ling-3.1-flash is already open-weight under MIT and it's big. Stats 🧠 ~560B total ⚡ ~25B active/token 📚 up to 1M context Community GGUFs already go down to IQ1_S → 142GB IQ2_XXS → 174GB IQ3_XXS → 256GB Q4_K_M → 354GB The existing Q4 has even been loaded across 4× DGX Sparks using llama.cpp RPC. ⚠️ 1019 is a new checkpoint/update, not a new Ling architecture announcement. Ant hasn't released the updated 1019 weights yet. It says open weights + API are coming soon. Link in ALT

2026.10.22 17:18

Wow, that's a pretty nice bump for Ling-3.1-flash. Ant introduced the new Ling-3.1-flash-1019 checkpoint, and its Artificial Analysis score jumped Ling-3.1-flash → 41 Ling-3.1-flash-1019 → 44 The highlighted gains come from its updated agentic performance AA-Briefcase ~1400 → 1524 Elo AutomationBench-AA 62% → 68% The existing Ling-3.1-flash is already open-weight under MIT and it's big. Stats 🧠 ~560B total ⚡ ~25B active/token 📚 up to 1M context Community GGUFs already go down to IQ1_S → 142GB IQ2_XXS → 174GB IQ3_XXS → 256GB Q4_K_M → 354GB The existing Q4 has even been loaded across 4× DGX Sparks using llama.cpp RPC. ⚠️ 1019 is a new checkpoint/update, not a new Ling architecture announcement. Ant hasn't released the updated 1019 weights yet. It says open weights + API are coming soon. Link in ALT

David Hendrickson@TeksEdge

2026.10.07 00:05

Ling-3.1-flash scores 41 on Artificial Analysis’s Intelligence Index versus its predecessor. The 560B model I posted last week now has another reason to stay on the watchlist. You can test it before the weights arrive on OpenCode because it has added free hosted access, and Vercel’s free trial runs through October 13. Vercel currently serves 262K context. The advertised ceiling is 1M. Still waiting on downloadable weights and confirmed local deployment requirements. Save the tasks it handles well now, then rerun them on a local quant when available. How much of that capability survives? Vercel link in ATL