DeepSeek V4.1-Flash-1017 claimed to hit Opus 5-level coding scores at 83x lower cost
Yuchenj_UW · x · 2026-10-18
X user Yuchen Jin claims the "OSS will never catch up" narrative collapsed within a week: a model dubbed DeepSeek-V4.1-Flash-1017 reportedly scores 77 on DeepSWE v1.1, matching Opus 5-level performance at 83x lower cost, with plans to launch it on Databricks.
Note this is a third-party claim — the model name and benchmark score are not yet confirmed by DeepSeek or independent evaluations.
More from Models
- Codex burns 20K credits in a day, $200-sub user calls out OpenAI over quota anomaly — alexcovo_eth · 2026-10-18
- Anthropic researcher predicts AI can harden world's cyber risk in 1-2 years; security engineers push back — joshua_saxe · 2026-10-18
- 3D watermelon test: Opus 5.5 costs $8.98 vs Gemini 4 Argon at $0.34 — DynamicWebPaige · 2026-10-18
- Prompting an AI to build and run a full 400-level economics course still doesn't work — rickasaurus · 2026-10-18
- Google touts Gemini 4 Argon finishing agentic task for $0.23 vs $33.45 on Claude — DynamicWebPaige · 2026-10-18
- DeepSeek V4.1 Flash claims to be ChatGPT in identity checks, screenshot shows — scaling01 · 2026-10-18