CharXiv-R

AI benchmark results and model performance

Type: percentageMultimodal

Recorded Results

Average Score

62.75%

Score Range

39.70% - 82.10%

Leading Model (lowest score)

39.70% - Qwen 3 VL 4B Instruct

Scores Over Time

Individual benchmark scores plotted by date.

Models Using This Benchmark

Organisation	Model	Reported	Top Score	Info	Self Reported	Source
Qwen	Qwen 3 VL 4B Instruct	-	39.70%	-	Yes	-
Qwen	Qwen 3 VL 8B Instruct	-	46.40%	-	Yes	-
Qwen	Qwen 3 VL 30B A3B Instruct	-	48.90%	-	Yes	-
Qwen	Qwen 3 4B Instruct 2507	-	50.30%	inferred family alias from qwen3-vl-4b-thinking (score=0.3462; benches=48)	Yes	-
Qwen	Qwen 3 4B SafeRL	-	50.30%	inferred family alias from qwen3-vl-4b-thinking (score=0.3850; benches=48)	Yes	-
Qwen	Qwen 3 4B Thinking 2507	-	50.30%	inferred family alias from qwen3-vl-4b-thinking (score=0.3462; benches=48)	Yes	-
Qwen	Qwen 3 Embedding 4B	-	50.30%	inferred family alias from qwen3-vl-4b-thinking (score=0.3850; benches=48)	Yes	-
Qwen	Qwen 3 Guard Gen 4B	-	50.30%	inferred family alias from qwen3-vl-4b-thinking (score=0.3400; benches=48)	Yes	-
Qwen	Qwen 3 Guard Stream 4B	-	50.30%	inferred family alias from qwen3-vl-4b-thinking (score=0.3371; benches=48)	Yes	-
Qwen	Qwen 3 Reranker 4B	-	50.30%	inferred family alias from qwen3-vl-4b-thinking (score=0.3850; benches=48)	Yes	-
Qwen	Qwen 3 VL 4B Thinking	-	50.30%	-	Yes	-
Qwen	Qwen 3 8B	-	53%	inferred high-confidence family alias from qwen3-vl-8b-thinking (score=0.4600; benches=50)	Yes	-
Qwen	Qwen 3 Guard Stream 8B	-	53%	inferred family alias from qwen3-vl-8b-thinking (score=0.3371; benches=50)	Yes	-
Qwen	Qwen 3 VL 8B Thinking	-	53%	-	Yes	-
Qwen	Qwen 3 VL Embedding 8B	-	53%	inferred high-confidence family alias from qwen3-vl-8b-thinking (score=0.5232; benches=50)	Yes	-
Qwen	Qwen 3 VL Reranker 8B	-	53%	inferred high-confidence family alias from qwen3-vl-8b-thinking (score=0.5275; benches=50)	Yes	-
OpenAI	GPT 4.5 Preview	27 Feb 2025	55.40%	-	Yes	Source
Qwen	Qwen 3 VL 30B A3B Thinking	-	56.60%	-	Yes	-
OpenAI	GPT 4o (2024-08-06)	06 Aug 2024	58.80%	-	Yes	Source
OpenAI	GPT 4o Audio Preview (2025-06-03)	03 Jun 2025	58.80%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
OpenAI	GPT 4o Audio Preview (2024-10-01)	01 Oct 2024	58.80%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
OpenAI	GPT 4o Audio Preview (2024-12-17)	17 Dec 2024	58.80%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
OpenAI	GPT 4o Realtime Preview (2025-06-03)	03 Jun 2025	58.80%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
OpenAI	GPT 4o Realtime Preview (2024-10-01)	01 Oct 2024	58.80%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
OpenAI	GPT 4o Search Preview	11 Mar 2025	58.80%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
OpenAI	GPT 4o Transcribe	20 Mar 2025	58.80%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
OpenAI	GPT 4o Transcribe Diarize	15 Oct 2025	58.80%	inferred modality/version alias from gpt-4o-2024-08-06	Yes	Source
Qwen	Qwen 3 VL 235B A22B Instruct	-	62.10%	-	Yes	-
Qwen	Qwen 3 VL 32B Instruct	-	62.80%	-	Yes	-
Qwen	Qwen 3 VL 32B Thinking	-	65.20%	-	Yes	-
Qwen	Qwen 3 VL 235B A22B Thinking	-	66.10%	-	Yes	-
OpenAI	o4 Mini	16 Apr 2025	72%	-	Yes	Source
OpenAI	o4 mini Deep Research	26 Jun 2025	72%	inferred modality/version alias from o4-mini	Yes	Source
Google	Gemini 3.1 Flash Image Preview (Nano Banana 2)	26 Feb 2026	73.20%	inferred modality/version alias from gemini-3.1-flash-lite-preview	Yes	Source
Google	Gemini 3.1 Flash-Lite	07 May 2026	73.20%	-	Yes	Source
Google	Gemini 3.1 Flash Lite Preview	03 Mar 2026	73.20%	-	Yes	Source
Google	Veo 3.1 Lite Preview	31 Mar 2026	73.20%	manual fallback alias from gemini-3.1-flash-lite-preview	Yes	Source
Qwen	Qwen 3.5 122B A10B	24 Feb 2026	77.20%	-	Yes	Source
Anthropic	Claude Opus 4.6	05 Feb 2026	77.40%	-	Yes	Source
Moonshot	Kimi K2.5	27 Jan 2026	77.50%	-	Yes	Source
Qwen	Qwen 3.5 35B A3B	24 Feb 2026	77.50%	-	Yes	Source
Qwen	Qwen 3.5 27B	24 Feb 2026	79.50%	-	Yes	Source
Qwen	Qwen 3.5 Flash	23 Feb 2026	79.50%	inferred family alias from qwen3.5-27b (score=0.4147; benches=81)	Yes	Source
Google	Gemini 3 Flash Preview	17 Dec 2025	80.30%	-	Yes	Source
Google	Gemini 3 Pro Image Preview (Nano Banana Pro)	20 Nov 2025	81.40%	inferred modality/version alias from gemini-3-pro-preview	Yes	Source
Google	Gemini 3 Pro Preview	18 Nov 2025	81.40%	-	Yes	Source
Qwen	Qwen 3.6 Plus	01 Apr 2026	81.50%	-	Yes	Source
OpenAI	GPT 5.2 Chat	11 Dec 2025	82.10%	inferred alias from gpt-5.2-2025-12-11	Yes	Source