General LLM SOTA: Gemini 4 Argon & O3 Mini & DeepSeek R1 & Llama 3.3…

isitsota / categories / General LLM

General LLM · Oct 13, 2026

Who is SOTA in General LLM?

Gemini 4 Argon & O3 Mini & DeepSeek R1 & Llama 3.3 70B Instruct & Gemma 3 27B IT & Phi 4

Best publicly available general-purpose language model, judged from independent preference and contamination-resistant composite evidence. As of Oct 13, 2026, Gemini 4 Argon & O3 Mini & DeepSeek R1 & Llama 3.3 70B Instruct & Gemma 3 27B IT & Phi 4 holds General LLM. On the cost-quality…

Overall Open weights ≤ 32B Best per $

Standings · pareto

The board

History →

#	Model	State	Confidence	Why
1	GGemini 4 Argon Google DeepMind	Pareto	medium	On the cost-quality frontier for General LLM: nothing scores better for less than $4.00 per 1M blended tokens.
4	OO3 Mini OpenAI	Pareto	medium	On the cost-quality frontier for General LLM: nothing scores better for less than $1.93 per 1M blended tokens.
5	OO1 OpenAI	Not SOTA · #5	medium	Not on the General LLM cost-quality frontier.
7	AClaude Opus 5.5 Anthropic	Not SOTA · #7	medium	Not on the General LLM cost-quality frontier.
7	DSDeepSeek R1 DeepSeek	Pareto	medium	On the cost-quality frontier for General LLM: nothing scores better for less than $1.15 per 1M blended tokens.
13	AClaude Opus 4.6 Anthropic	Not SOTA · #13	medium	Not on the General LLM cost-quality frontier.
19	AClaude Fable 5 Anthropic	Not SOTA · #19	medium	Not on the General LLM cost-quality frontier.
25	AClaude Opus 4.7 Anthropic	Not SOTA · #25	medium	Not on the General LLM cost-quality frontier.
28	OGPT-4O OpenAI	Not SOTA · #28	medium	Not on the General LLM cost-quality frontier.
29	MLlama 3.3 70B Instruct Meta	Pareto	medium	On the cost-quality frontier for General LLM: nothing scores better for less than $0.64 per 1M blended tokens.
30	GGemma 3 27B IT Google DeepMind	Pareto	medium	On the cost-quality frontier for General LLM: nothing scores better for less than $0.17 per 1M blended tokens.
31	AClaude Fable 5.1 Anthropic	Not SOTA · #31	medium	Not on the General LLM cost-quality frontier.
37	GGemini 3.8 Flash Google DeepMind	Not SOTA · #37	medium	Not on the General LLM cost-quality frontier.
42	MSPhi 4 Microsoft	Pareto	medium	On the cost-quality frontier for General LLM: nothing scores better for less than $0.09 per 1M blended tokens.
43	OGPT-4O Mini OpenAI	Not SOTA · #43	medium	Not on the General LLM cost-quality frontier.

Pareto

Quality against blended price

Evidence recipe

What can move this crown

LMArena Text (style-controlled, overall) — primary, weight 1
Artificial Analysis Intelligence Index — primary, weight 1
LiveBench — secondary, weight 0.45