Skip to content

Benchmark

Humanity's Last Exam

Scores are percentages. Higher is better.

Models and systems measured
120
Scores recorded
230
Sources
3
IQ weight
1.67

Score distribution

34%
Mean
34.8%
Median
15.5%
Std dev
12.5%
10th percentile
55.7%
90th percentile
3.7%
Min
3.7%73.2%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.67 for chat models, measured over 113 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#Model or systemScore
151Qwen3.6 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim28.8%
152Qwen3.6 PlusQwen official release table · thinking mode unless benchmark requires otherwise· Reported by Alibaba Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.28.8%
153Qwen3.5 397B A17B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
28.7%
154Qwen3.5 397B A17BQwen3.5 release/model card · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.28.7%
155GPT-5.4 mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim28.2%
156GPT-5.4 minixhigh · without tools· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.28.2%
157Step 3.5 FlashStep 3.5 technical report · Tables 12–13 · text-only · Python tool + PaCoRe· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.28.2%
158Step 3.5 FlashStep 3.5 technical report · Table 5 · text-only · PaCoRe· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.27.9%
159Nemotron 3 UltraNeMo Evaluator SDK · no tools· Reported by NVIDIA· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.26.7%
160Gemma 4 31B IT· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim26.5%
161Step 3.5 FlashStep 3.5 technical report · Tables 12–13 · text-only · Python tool· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.26.5%
162LongCat-Flash-Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
25.2%
163DeepSeek Reasoner· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim25.1%
164DeepSeek V3.2Thinking mode · temperature 1.0 · 128K context · custom step-by-step / boxed-answer prompt· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.25.1%
165GPT 5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim24.8%
166GPT 5.4 Nano· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim24.3%
167GPT 5.4 Nanoxhigh · without tools· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.24.3%
168Qwen3.6 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim24%
169Qwen3.6 27BQwen official release table · thinking mode unless benchmark requires otherwise· Reported by Alibaba Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.24%
170DeepSeek V3.2Thinking mode · temperature 1.0 · 128K context · official HLE template· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.23.9%
171Step 3.5 Flash· OpenEvals Leaderboard data· Reported by the model's developer. The source relaying it did not verify it.23.1%
172Step 3.5 FlashStep 3.5 technical report · Table 5 · text-only · vanilla· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.23.1%
173Nemotron 3 Super 120B A12B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
22.8%
174MiniMax M2.1· OpenEvals Leaderboard data· Reported by the model's developer. The source relaying it did not verify it.
Values from 2 sources
22.2%
175MiMo V2 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
22.1%
176MiMo V2 FlashMiMo-V2-Flash final evaluation · text-only · no tools· Reported by Xiaomi MiMo· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.22.1%
177Muse Glimmer 30B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim22%
178Muse Glimmer 30BText · AA · high reasoning· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.22%
179Gemini 2.5 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim21.6%
180Qwen3.6 35B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim21.4%
181Qwen3.6 FlashQwen3.6-35B-A3B official model card · hosted API name Qwen3.6-Flash· Reported by Alibaba Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.21.4%
182Kimi K2.5image subset · no tools· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.21.3%
183Grok 4 Fast· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim20%
184DeepSeek V3.2 Exp· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim19.8%
185Gemma 4 31B ITGemma 4 official model card · no tools· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.19.5%
186MiniMax M2.5· OpenEvals Leaderboard data· Reported by the model's developer. The source relaying it did not verify it.19.4%
187MiniMax M2.5official appendix · no tools· Reported by MiniMax· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.19.4%
188Nemotron 3 Super 120B A12BNeMo Evaluator / mostly NeMo Skills · no tools· Reported by NVIDIA· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.18.3%
189Qwen 3 235B A22B Thinking 2507· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim18.2%
190Claude Sonnet 4.5reasoning only · without search/tools · decontaminated · Sonnet 4.5 grader· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.17.7%
191DeepSeek R1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim17.7%
192Gemma 4 26B A4B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim17.2%
193GLM 4.6· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim17.2%
194GPT-5 mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim16.7%
195Gemini 3.1 Flash Lite· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim16%
196Gemini 3.1 Flash LiteGemini 3.1 Flash-Lite Model Card · high thinking · default sampling · single attempt/pass@1 · no tools· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.16%
197DeepSeek V3.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim15.9%
198Nemotron 3 Nano 30B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
15.5%
199GPT OSS 120B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
14.9%
200GPT OSS 120Bofficial model card Table 3 · reasoning high · no tools · pass@1· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.14.9%

Showing 151–200 of 229 results.

Not ranked

The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.

Model or systemScore
Muse Spark (version unspecified)Meta Model API · highest effort · version unspecified · no tools · full 2,500 questions · o3-mini judge· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.42.8%

Data credits: Data by LLM Stats · BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links. · OpenEvals/leaderboard-data, MIT (licence)