⌕Is ___ SOTA?
isitsota / categories / Factuality
Also in Text:General LLMReasoningMath+2C
Factuality · Oct 14, 2026
Who is SOTA in Factuality?
Gemini 4 Argon & Mimo V2.6 Pro & DeepSeek V4.1 Flash & GPT-6 Luna & Mimo V2.6 Flash & DeepSeek V4 Flash & GPT Oss 120B

Short-form factual accuracy. SimpleQA Verified plus the LMArena factuality slice. As of Oct 14, 2026, Gemini 4 Argon & Mimo V2.6 Pro & DeepSeek V4.1 Flash & GPT-6 Luna & Mimo V2.6 Flash & DeepSeek V4 Flash & GPT Oss 120B holds…

OverallOpen weights≤ 32BBest per $
Standings · pareto

The board

RSSHistory →
#ModelStateConfidenceWhy
1GGemini 4 Argon
Google DeepMind		Pareto · FactualitymediumOn the cost-quality frontier for Factuality: nothing scores better for less than $4.00 per 1M blended tokens.
8XIMimo V2.6 Pro
xiaomi		Pareto · FactualitymediumOn the cost-quality frontier for Factuality: nothing scores better for less than $0.54 per 1M blended tokens.
25DSDeepSeek V4.1 Flash
DeepSeek		Pareto · FactualitymediumOn the cost-quality frontier for Factuality: nothing scores better for less than $0.22 per 1M blended tokens.
45OGPT-6 Luna
OpenAI		Pareto · FactualitymediumOn the cost-quality frontier for Factuality: nothing scores better for less than $0.20 per 1M blended tokens.
50XIMimo V2.6 Flash
xiaomi		Pareto · FactualitymediumOn the cost-quality frontier for Factuality: nothing scores better for less than $0.18 per 1M blended tokens.
84DSDeepSeek V4 Flash
DeepSeek		Pareto · FactualitymediumOn the cost-quality frontier for Factuality: nothing scores better for less than $0.11 per 1M blended tokens.
150OGPT Oss 120B
OpenAI		Pareto · FactualitymediumOn the cost-quality frontier for Factuality: nothing scores better for less than $0.07 per 1M blended tokens.
2AClaude Fable 5.1
Anthropic		Near-SOTA · FactualitymediumNot on the Factuality cost-quality frontier.
3AClaude Opus 4.6
Anthropic		Near-SOTA · FactualitymediumNot on the Factuality cost-quality frontier.
4AClaude Opus 5
Anthropic		Near-SOTA · FactualitymediumNot on the Factuality cost-quality frontier.
5OGPT-5.5
OpenAI		Near-SOTA · FactualitymediumNot on the Factuality cost-quality frontier.
1AClaude Opus 5.5
Anthropic		#1 in FactualitymediumNot on the Factuality cost-quality frontier.
1MMuse Wonder
Meta		#1 in FactualitymediumNot on the Factuality cost-quality frontier.
6OGPT-5.4
OpenAI		#6 in FactualitymediumNot on the Factuality cost-quality frontier.
7AClaude Fable 5
Anthropic		#7 in FactualitymediumNot on the Factuality cost-quality frontier.
Pareto

Quality against blended price

Gemini 4 ArgonMimo V2.6 ProDeepSeek V4.1 FlashGPT-6 LunaMimo V2.6 FlashDeepSeek V4 FlashGPT Oss 120BClaude Fable 5.1Claude Opus 4.6Claude Opus 5GPT-5.5Claude Opus 5.5
Evidence recipe

What can move this crown