DeepSeek V4.1 Flash hands-on tests and comparisons

Per DeepSeek's own chart, Flash costs about $0.20 per task and scores 77.0 on DeepSWE vs Opus 5's 74.0, though Opus still leads Terminal-Bench 51.8 to 42.3.

The post says it edges past Opus 5 on DeepSWE, ships MIT open weights, is live on OpenRouter, keeps the prior Flash price, and has a surprisingly low cache price.

DeepSeek calls it a Pareto improvement on performance and cost: DeepSWE 77.0 above Opus 5's 74.0 and GPT 5.6-Sol's 73.0, about $0.2 per task at max tier, with Harness and CodeBuddy adaptations.

DeepSeek's new Flash model is live across web, app and API with open weights, hitting 77% on DeepSWE at about 1.3 yuan per task and improving all 11 benchmarks.

A Japanese post says V4.1 Flash appears post-trained for Dsh, scores above Astra Fable, ships open weights on Hugging Face, and shows strong 3D CG demos.

Official notice says to call deepseek-flash, now upgraded to V4.1-Flash-1017 with unchanged usage; legacy names like deepseek-v4-flash still work but are served by the 1017 model at Flash pricing.

V4.1-Flash-1017 ships MIT-licensed weights with the same ~550B/8B-16B-active architecture, Terminal-Bench up 31 to 42 and DeepSWE 77 above Opus 5; the checkpoint is about 510GB.

Terminal-Bench 42.3 in the Claude Code harness, DeepSWE 77.0, 1M context and MIT weights, with $0.60/1M output off-peak plus step-by-step Claude Code setup and free Search/Fetch until Nov 1.

User reports that V4.1 Flash 1017, trained with self-feedback RL, improves building complete projects from scratch, especially game development and 3D modeling tasks.

The official release says the new model is tuned for Harness with faster tasks, shorter outputs and lower token costs; users need 0.2.1-rc.1 or later, and first-time users get a 6-yuan credit.

A user recap: gains in game dev and 3D modeling, shorter task time and outputs, a web search API indexing hundreds of billions of pages, and a recommendation to use DeepSeek Harness.

After the official DeepSeek V4.1 Flash release, a user is running their own tests and inviting others to send prompts to try.