October model tier board places DeepSeek V4.1-Flash-1017 in A tier at $0.30/$1.20
A third-party October 2026 tier list puts DeepSeek V4.1-Flash-1017 in A tier alongside GPT-6.1 Sol and Sonnet 5.5, below S+ models Opus 5.5, Gemini 4 Argon and Fable 5, and lists its $0.30/$1.20 pricing.
Updated the board for October 2026 after testing this week's drops: - S+: Claude Opus 5.5, Gemini 4 Argon ($2/$10), Fable 5 - A: GPT-6.1 Sol, DeepSeek V4.1-Flash-1017 ($0.30/$1.20), Sonnet 5.5, Opus 5 - B: GPT-5.6 Sol, Gemini 3.8 Flash, GLM-5.3, Muse Spark Open Which placement on here are you arguing with first?
And the token math gets even crazier on weekends/off-peak when V4.1-Flash-1017 drops in half to $0.15 / $0.60, that's ~16x cheaper than Argon's $2/$10 intro rate (and 33x cheaper once Argon doubles to $4/$20 on Dec 1), while scoring 77.0% vs Argon's 77.9% on DeepSWE v1.1. For high-volume sub-agents, Flash 1017 is untouchable right now.