Leaked test code hints at DeepSeek V4 Mini with 1M-token context and a V4 Pro sibling
teortaxesTex · x · 2026-09-30
A leak from DeepSeek's Harness repo test code suggests a small model is coming:
- Model ID deepseek-v4-mini, display name DeepSeek-V4-Flash, with context hard-coded at 1,000,000 tokens
- The same list also shows deepseek-v4-pro described as "Preserved hidden detail"
- Frontend code already handles model switching and Apply/validation flows—less like a throwaway mock, more like a lighter desktop-facing model in the works
teortaxesTex notes a single-GPU, 1M-context, heavily post-trained model with agent teams, local-first skills, trained on Huawei Ascend and with the recipe published could be an "R1 moment #2". Unconfirmed leak.
More from Models
- Stealth Anthropic model spotted in Devin build as METR chart shows frontier pace up 1.4x — daniel_mac8 · 2026-11-16
- Eval Firm Trajectory Told 40 Agents "NEVER Stop" in Harness Without Telling Epoch — peterwildeford · 2026-11-16
- METR time-horizon chart updated: doubling every 1.7 months since Opus 4.5, 'Model 4' ~38h — ben_j_todd · 2026-11-16
- Gemini Pro user says its coding agent rejects every Three.js/SVG prompt as 'illegal' — Necessary-Buy-7410 · 2026-11-16
- 28-model tracking table: half publish no training cutoff, Gemini 4 Carbon 8 months stale at launch — Joozio · 2026-11-16
- Opus 5.5 caught using account email to identify user in supposedly memory-less session — birchlse · 2026-11-16