Tau Bench (Retail)

Overview

Type: percentage

Recorded Results

Average Score

61.09%

Score Range

21.70% - 82.40%

Leading Model (lowest score)

21.70% - EXAONE 4.0 1.2B

Scores Over Time

Individual benchmark scores plotted by date.

Models Using This Benchmark

Organisation	Model	Reported	Top Score	Info	Self Reported	Source
LG	EXAONE 4.0 1.2B	15 Jul 2025	28.10%	Reasoning	Yes	Source
OpenAI	GPT OSS 20b	05 Aug 2025	54.80%	High Reasoning Effort	Yes	Source
OpenAI	GPT OSS Safeguard 20b	29 Oct 2025	54.80%	inferred high-confidence family alias from gpt-oss-20b (score=0.5137; benches=7)	Yes	Source
OpenAI	o3 mini	30 Jan 2025	57.60%	-	Yes	Source
OpenAI	GPT 4o (2024-08-06)	06 Aug 2024	60.30%	-	Yes	Source
OpenAI	GPT 4o Audio (2025-06-03)	03 Jun 2025	60.30%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
OpenAI	GPT 4o Audio (2024-10-01)	01 Oct 2024	60.30%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
OpenAI	GPT 4o Audio (2024-12-17)	17 Dec 2024	60.30%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
OpenAI	GPT 4o Realtime Preview (2025-06-03)	03 Jun 2025	60.30%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
OpenAI	GPT 4o Realtime Preview (2024-10-01)	01 Oct 2024	60.30%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
OpenAI	GPT 4o Search Preview	11 Mar 2025	60.30%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
OpenAI	GPT 4o Transcribe	20 Mar 2025	60.30%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
OpenAI	GPT 4o Transcribe Diarize	15 Oct 2025	60.30%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
Qwen	Qwen 3 Next 80B A3B Instruct	-	60.90%	TAU1-Retail	Yes	Source
LG	EXAONE 4.0 32B	15 Jul 2025	62.80%	Reasoning	Yes	Source
MiniMax	MiniMax M1 80K	16 Jun 2025	63.50%	-	Yes	-
MiniMax	Minimax M1 40K	16 Jun 2025	67.80%	-	Yes	-
OpenAI	GPT OSS 120b	05 Aug 2025	67.80%	High Reasoning Effort	Yes	Source
OpenAI	GPT OSS Safeguard 120b	29 Oct 2025	67.80%	inferred high-confidence family alias from gpt-oss-120b (score=0.5102; benches=7)	Yes	Source
Qwen	Qwen 3 235B A22B Thinking 2507	-	67.80%	TAU1-Retail	Yes	Source
OpenAI	GPT 4.5	27 Feb 2025	68.40%	-	Yes	Source
Qwen	Qwen 3 Next 80B A3B Thinking	-	69.60%	TAU1-Retail	Yes	Source
Qwen	Qwen 3 A235 A22B Instruct 2507	-	71.30%	-	Yes	Source
OpenAI	o4 Mini	16 Apr 2025	71.80%	-	Yes	Source
OpenAI	o4 mini Deep Research	26 Jun 2025	71.80%	inferred modality/version alias from o4-mini	Yes	Source
Qwen	Qwen 3 Coder 480B A35B Instruct	-	77.50%	-	Yes	Source
z.AI	GLM 4.5 Air	28 Jul 2025	77.90%	-	Yes	Source
z.AI	GLM 4.5	28 Jul 2025	79.70%	-	Yes	Source
Anthropic	Claude Opus 4.1	05 Aug 2025	82.40%	-	Yes	Source