← October 17, 2026 briefing · added in the 20:00 KST update

DeepSeek releases 'V4.1-Flash-1017' on Hugging Face; Terminal-Bench 4.0 rises from 31.2 to 42.3 (community reports)

Announced October 17, 2026

Verified (fact)

According to community reports, DeepSeek has released 'DeepSeek-V4.1-Flash-1017', a minor update to V4.1 Flash, on Hugging Face. Earlier, there were also reports that 'deepseek-v4.1-flash-1017' briefly appeared in the API model list before being taken down. Shared public evaluation results show Terminal-Bench 4.0 rising from 31.2 to 42.3 and ProgramBench from 20.3 to 29.1. Performance on reasoning, coding, agent tasks and vision benchmarks reportedly improved. On the same day, the agent tool 'DeepSeek Harness' v0.2.1-rc1 was also released. This version adds user-specified plugin indexes, plugins that agents create on their own, and reasoning settings that can be configured per model.

Needs fact-check (claim)

The benchmark figures (Terminal-Bench 4.0 42.3, ProgramBench 29.1) come from the update announcement and have not been independently verified.

Why it matters (interpretation)

The Flash model, already widely used (OpenCode offers it for free, for example), was updated just a week after release. With agentic coding scores up sharply, teams running agents on low-cost open models may want to consider switching.

Confidence medium · official source pending

Sources

Story thread

  1. 2026-10-17 DeepSeek releases 'V4.1-Flash-1017' on Hugging Face; Terminal-Bench 4.0 rises from 31.2 to 42.3 (community reports)
  2. 2026-10-17 DeepSeek V4.1 Flash 1017 follow-up: company's first CED (Causal Encoder-Decoder) architecture, native image understanding; 552B total, 8B input/16B output active (community reports)
  3. 2026-10-18 DeepSeek V4.1-Flash-1017 follow-up: official release confirmed with MIT-licensed weights, rolled out to web, app and API at unchanged prices (report)
  4. 2026-10-19 DeepSeek V4.1-Flash-1017 follow-up: independent Terminal-Bench 4.0 scores of 34% and 31.8% trail the model card's 42.3% (community reports)

More about DeepSeek