DeepSeek V4.1 Flash becomes first open model to crack top 10 on search leaderboard
gaganghotra_ · x · 2026-10-19
DeepSeek released DeepSeek-V4.1-Flash-1017, described as a Pareto improvement over the previous version: stronger benchmark performance, faster task completion, and shorter outputs. It's now live on the website, app, and API.
Third-party evaluator p0 reports it's the highest-performing open model they've measured, and the first open model to crack the top 10 on the Search Capability Leaderboard — the chart shows it at #10 with 68.9 (+36.0 from search), alongside closed frontier models like GPT-6.1 Sol (70.4) and Claude Opus 5 (70.0).
The model is optimized for DeepSeek Harness, Tencent CodeBuddy, and WorkBuddy; DeepSeek recommends Harness v0.2.1-rc.1+ for best results.
More from coding & agent
- Indie dev builds Utsushi, an open-source Higgsfield clone in Go/HTMX — _playlogic_ · 2026-10-19
- Dust hyper-optimization delivers ~10x speedup, verified; beyond-backprop algorithms coming — burny_tech · 2026-10-19
- Writer runs all-day workflow in Claude's in-app browser, agents apply doc edits as suggestions — danshipper · 2026-10-19
- 30 Days, 30 AI Agents: A Daily Series on AI Agents for Biotech and Pharma R&D — nlarusstone · 2026-10-19
- AI assistant caught recommending its user's portfolio company, then re-justifies the pick on merit — matt_slotnick · 2026-10-19
- Hark agent demos auto-filing a DMV address change in 2 minutes, no browser — adcock_brett · 2026-10-19