DeepSeek released V4.1-Flash-1017 with MIT license and new Search and Fetch API tools; Databricks listed it as an endpoint on the same day.
Following·New·2 sources
DeepSeek V4.1 Flash 1017 launches open-weight, beats Opus 5 on coding benchmark
The gist
The model's MIT license allows broad commercial use at low cost. The pricing gap between DeepSeek's API and comparable offerings is substantial.
The full picture
DeepSeek released V4.1-Flash-1017, an open-weight model update available for free download under an MIT license. On the DeepSWE v1.1 coding benchmark, it scored 77.0, compared to 74.0 for Claude Opus 5 and 73.0 for GPT-5.6 Sol, though DeepSeek ran its own tests. Via DeepSeek's API, pricing is $0.15 per million input tokens and $0.60 per million output tokens off-peak, versus roughly $1.86 for GPT-5.6 Sol and $3.15 for Claude Fable 5 on a standard test suite. DeepSeek also launched Search and Fetch API tools at $0.75 per 1,000 requests, compared to rivals Brave, Tavily, and Perplexity at $5 per 1,000 and up. Free usage through DeepSeek Harness, Claude Code, or Codex ends November 1. Databricks made the model available as an endpoint on the same day it shipped, routing all calls through Unity Gateway for unified access and spend monitoring.
How it developed
Sources
The daily email
Want this in your inbox?
I send one email each morning with the stories that moved. If you would rather just read here, that works too.
Subscribe free