Skip to content

Benchmark

Vending-Bench 2

Agentic benchmarks measure how well models carry out tasks that take several steps. Higher scores are better.

Models measured
69
Scores recorded
87
Sources
3
IQ weight
2.11

Score distribution

6,546.7
Mean
5,165
Median
8,742.6
Std dev
458.9
10th percentile
11,181.9
90th percentile
-31.2
Min
-31.280,017.6
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.11 for chat models, measured over 69 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Opus 4.6Anthropic Opus 4.6 launch page · adaptive thinking · max effort unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.80,017.6
2Gemini 4 Neon· Epoch AI Benchmarking Hub data16,843.4
3Gemini 4 Carbon· Epoch AI Benchmarking Hub data
Values from 2 sources
16,484.2
4GPT-6 Astra· Epoch AI Benchmarking Hub data15,514.7
5Gemini 4 Hydrogen· Epoch AI Benchmarking Hub data14,549.5
6GPT-6 Sol· Epoch AI Benchmarking Hub data14,427.8
7Gemini 4 Argon· Epoch AI Benchmarking Hub data13,718.2
8GPT-6.1 Sol· Epoch AI Benchmarking Hub data12,860.6
9Claude Opus 5· Epoch AI Benchmarking Hub data11,181.9
10Claude Sonnet 5.5· Epoch AI Benchmarking Hub data11,165.3
11Claude Opus 4.7final balance · max effort · benchmark-native context management· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.10,937
12Claude Opus 4.7· Epoch AI Benchmarking Hub data10,936.8
13Grok 4.7· Epoch AI Benchmarking Hub data10,536.8
14Claude Fable 5.2· Epoch AI Benchmarking Hub data9,628.5
15GPT-5.6 Sol· Epoch AI Benchmarking Hub data9,619.4
16Muse Wonder 1.0· Epoch AI Benchmarking Hub data9,432.2
17Claude Opus 5.5· Epoch AI Benchmarking Hub data9,235.2
18GPT-6.1 Terra· Epoch AI Benchmarking Hub data9,107.5
19Grok 4.6· Epoch AI Benchmarking Hub data9,047
20GLM 5.2· Epoch AI Benchmarking Hub data8,313.8
21GLM 5.3· Epoch AI Benchmarking Hub data8,163.6
22Claude Opus 4.6· Epoch AI Benchmarking Hub data
Values from 2 sources
8,017.6
23Claude Opus 4.7final balance · high effort · benchmark-native context management· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.7,971
24GPT 5.5· Epoch AI Benchmarking Hub data7,523.8
25GPT-5.6 Terra· Epoch AI Benchmarking Hub data7,343.2
26Claude Sonnet 4.6final balance · max effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.7,204.1
27Claude Sonnet 4.6· Epoch AI Benchmarking Hub data7,204.1
28Claude Sonnet 4.6final balance · high effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.6,625.1
29Muse Spark 1.1· Epoch AI Benchmarking Hub data6,520.5
30Claude Sonnet 5· Epoch AI Benchmarking Hub data6,377.7
31Kimi K2.6· Epoch AI Benchmarking Hub data6,204.6
32GPT 5.4· Epoch AI Benchmarking Hub data6,144.2
33GPT-5.3 Codex· Epoch AI Benchmarking Hub data5,940.1
34Claude Opus 4.8· Epoch AI Benchmarking Hub data5,787.4
35Claude Fable 5high· Epoch AI Benchmarking Hub data5,680.3
36GLM 5.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
5,634.4
37GLM 5.1GLM-5.1 official release table · average net assets· Reported by Z.ai· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.5,634.4
38Gemini 3 Pro Preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
5,478.2
39Gemini 3 Pro PreviewGemini 3 Pro Model Card · results as of November 2025 · mean net worth· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.5,478.2
40Claude Fable 5.1· Epoch AI Benchmarking Hub data5,421.6
41Gemini 3.5 Flash· Epoch AI Benchmarking Hub data5,396.4
42Kimi K3· Epoch AI Benchmarking Hub data5,165
43Qwen3.6 Plus· Epoch AI Benchmarking Hub data5,114.9
44Gemini 3.8 Flash· Epoch AI Benchmarking Hub data5,093.8
45Kimi K2.7 Code· Epoch AI Benchmarking Hub data5,082.9
46Claude Fable 5low· Epoch AI Benchmarking Hub data5,018.5
47Claude Opus 4.5final balance · high effort · 8,192 reasoning tokens/turn · benchmark-native context management· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.4,967.1
48Claude Opus 4.5· Epoch AI Benchmarking Hub data4,967.1
49Claude Fable 5max· Epoch AI Benchmarking Hub data4,966.6
50Grok 4.20· Epoch AI Benchmarking Hub data4,662.8

Showing 1–50 of 87 results.

Data credits: Epoch AI Benchmarking Hub, CC BY 4.0 (licence) · 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. · Data by LLM Stats