Loading activity…
@hellosurgeai
Our mission is to raise AGI with the richness of humanity — curious, witty, imaginative, and full of breathtaking brilliance.
AI
Surge AI also points to legacy code nobody wants to touch.
AI
Mistral says Mistral Large 4 is live via API in public preview now, with open weights due at the end of October. Reuters says the launch is part of Mistral’s push to narrow the gap with leading Chinese open-weight models.
AI
Surge AI says it post-trained Kimi K2.7 on 1,465 GDP.pdf companion tasks. Its GDP.pdf score rose from 11.0% to 24.5%, while full-task success on GDPval more than doubled.
AI
HelloSurgeAI says DAYJOB contains 130 expert-built assignments modeled on real professional work. As of September 28, it said the strongest models succeeded on 24.7% of Healthcare assignments and 23.9% of Finance assignments.
AI
Surge AI says the model's additional training covered documents, spreadsheets, research, planning and tool use—with zero coding tasks. It attributes the coding benchmark gain to more general task-execution skills.
AI
Surge AI says training Qwen3-4B on 1,000 ComplexConstraints examples raised its rubric pass rate from 57.9% to 73.4%, close to Qwen3-235B-A22B-Instruct’s 73.9%.
AI
Surge AI, which ran the human baseline, notes a key caveat: the agents could iterate repeatedly, while each of the 28 human researchers submitted one idea.
