Skip to content

Benchmark

Arena ELO Text Factuality - Instruction Following

Instruction following benchmarks measure how closely models follow instructions. Higher scores are better.

Models measured
160
Scores recorded
179
Sources
1
IQ weight
2.15

Score distribution

1,436.3
Mean
1,439.3
Median
44.1
Std dev
1,377.3
10th percentile
1,486.3
90th percentile
1,280.8
Min
1,280.81,523.5
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.15 for chat models, measured over 160 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
101Qwen3.6 Plus· LMArena data1,434.3
102Grok 4.7xHigh· LMArena data1,433.7
103GPT-5.5 Instant· LMArena data1,433.6
104Gemini 3 FlashMinimal· LMArena data1,432.8
105DeepSeek-V4-Flash PreviewHigh· LMArena data1,432.2
106Grok 4.20 Beta 1· LMArena data1,431.9
107DeepSeek V3.2Thinking· LMArena data1,431.7
108Inkling 2 Small· LMArena data1,431.1
109DeepSeek V3.2· LMArena data1,430
110Mistral Medium 3.5· LMArena data1,429.6
111Claude Haiku 5.5High· LMArena data1,429.3
112Claude Haiku 4.520251001· LMArena data1,428.2
113ChatGPT-4o Latest20250326· LMArena data1,427.5
114ERNIE 5.00110· LMArena data1,426.5
115NVIDIA Nemotron 3.5 Super VL 120BNVFP4· LMArena data1,425.4
116Qwen3 235B A22B Instruct 25072507· LMArena data1,425.1
117Gemini 2.5 Flash· LMArena data1,425.1
118Mistral Large 4· LMArena data1,424.7
119DeepSeek V3.2 Exp· LMArena data1,424.1
120MiniMax M2.7· LMArena data1,424
121Nemotron 3 UltraNVFP4· LMArena data1,423.5
122GLM 4.6· LMArena data1,423.4
123GLM 4.7· LMArena data1,423.1
124Mistral Large 3· LMArena data1,422.5
125LongCat-Flash-Chat Exp2602· LMArena data1,422.3
126Inkling Small· LMArena data1,421.3
127Gemini 2.5 Flash Preview09-2025· LMArena data1,417
128ERNIE-5.0-Preview1203· LMArena data1,417
129MiniMax-M2.1 Preview· LMArena data1,416.9
130Qwen3 Max Preview· LMArena data1,414.1
131Kimi K2 Thinking· LMArena data1,414
132Qwen3.5 27B· LMArena data1,413.6
133Qwen3.5 122B-A10B· LMArena data1,412.9
134Mistral Medium 3.12508· LMArena data1,412.8
135Grok 4.1Thinking· LMArena data1,412.3
136Doubao Seed 2.0 Pro· LMArena data1,411.6
137MiMo V2 FlashNon-Thinking· LMArena data1,410.6
138Step 3.5 Flash· LMArena data1,409.9
139Grok 4.1· LMArena data1,408.7
140Gemini 3.1 Flash Lite· LMArena data1,407
141Grok 4 FastReasoning· LMArena data1,406.5
142Grok 4.3· LMArena data1,406.3
143GPT 5High· LMArena data1,406.1
144GPT-5.3 Chat· LMArena data1,405.4
145Solar Pro 4· LMArena data1,405.1
146GPT 5.4 NanoHigh· LMArena data1,402.8
147Qwen3.5 Flash· LMArena data1,400.9
148GPT-5 Chat· LMArena data1,400.8
149Grok 40709· LMArena data1,400.1
150Qwen3.5 35B A3B· LMArena data1,400

Showing 101–150 of 179 results.

Data credits: LMArena, CC BY 4.0 (licence)