Overview
"DeepSeek Flash 4.1 1017" refers to DeepSeek-V4.1-Flash-1017, the October 17, 2026 update served under the API name deepseek-flash. MiniMax M3.1 Flash launched two days later, on October 19, 2026.
Pricing
DeepSeek V4.1-Flash-1017: $0.30 per 1M cache-miss input tokens and $1.20 per 1M output tokens at peak; off-peak rates are half that ($0.15 / $0.60).MiniMax M3.1 Flash: a flat $0.15 per 1M input tokens, $0.60 per 1M output tokens and $0.015 per 1M cached input tokens.Off-peak DeepSeek and MiniMax charge the same for uncached tokens. At peak, DeepSeek costs twice as much for uncached tokens but less for cached input.Benchmarks (as reported by each vendor)
Both companies report DeepSWE v1.1: 77.0 for DeepSeek V4.1-Flash-1017 and 73.7% for MiniMax M3.1 Flash. Test harnesses may differ, since DeepSeek says it used the Mini-SWE-V2 framework.Both also report Agents' Last Exam: 35.4 for DeepSeek and 29.6 for MiniMax.DeepSeek also reports Terminal-Bench 4.0 at 42.3 and HLE without tools at 42.6.MiniMax also reports OSWorld V2 at 61.7, CyberGym at 82% and BankerToolBench at 76.2, plus output speeds above 200 tokens per second with a median of 170+.Context and availability
DeepSeek lists a 1M-token context and 384K max output, with vision support.MiniMax's launch post does not state M3.1 Flash's context length. It is available through the MiniMax API, M Plan and MiniMax Code, and MiniMax says the weights will be released later.No hands-on testing was conducted for this comparison. Every figure above comes from the official pages listed under Research sources, as published on or before October 22, 2026. Where a vendor has not published a number, we say so instead of estimating it.