Leaderboards
Benchmarks
20+
Including benchmarks on agentic coding, frontier reasoning, and safety alignment.
Models evaluated
100+
From leading AI labs including OpenAI, Anthropic, Google, Meta, and open-source contributors.
HiL-Bench tests whether agents know when to ask for help, measuring if they recognize missing or ambiguous information and ask targeted clarifying questions instead of guessing.
Claude Opus 4.7
27.67±5.32
Claude Opus 4.6
24.33±5.16
GLM-5.1
21.00±4.96
Evaluating an agent's ability to write production-grade tests
Gpt-5.4-xHigh (Codex CLI)
44.36±6.04
Gpt-5.4-xhigh (Mini-SWE)
40.00±6.00
Gpt-5.3-Codex-Xhigh (Codex)
38.98±6.12
Evaluating deep code comprehension and reasoning
Gpt 5.4 xHigh (Codex)
40.80±5.10
Gpt 5.4 xHigh (Mini-SWE-Agent)
36.30±4.90
Opus 4.6 (Claude Code)
33.30±5.00
Evaluating real-world tool use through the Model Context Protocol (MCP)
Muse Spark
82.20±2.30
claude-opus-4-7 (max)
79.10±2.50
gemini-3.1-pro-preview (high)
78.20±2.50
Evaluating long-horizon software engineering tasks in public open-source repositories
gpt-5.4 (xHigh)*
59.10±3.56
Muse Spark*
New55.00±3.60
claude-opus-4-6 (thinking)*
51.90±3.61
Evaluating long-horizon software engineering tasks in commercial-grade private repositories
claude-opus-4-6 (thinking)*
47.10±6.07
Muse Spark*
New44.70±6.05
gpt-5.4(xHigh)*
43.40±6.03
Forecasting scientific experiment outcomes
gemini-3-pro-preview
25.27±1.92
claude-opus-4-5-20251101
23.05±0.51
claude-opus-4-1-20250805
22.22±1.48
Challenging LLMs at the frontier of human knowledge
gemini-3.1-pro-preview (thinking high)
46.44±1.96
gpt-5.4-pro-2026-03-05
44.32±1.95
Muse Spark
New40.56±1.92
Challenging LLMs at the frontier of human knowledge
gemini-3.1-pro-preview (thinking high)
47.31±2.11
gpt-5.4-pro-2026-03-05
45.32±2.10
Muse Spark
New40.92±2.07
Evaluating spoken dialogue systems in multi-turn interaction
gemini-3-pro-preview (Thinking)*
54.65±4.57
gemini-2.5-pro (Thinking)*
46.90±4.58
gemini-2.5-flash (Thinking)*
40.04±4.50
Evaluating spoken dialogue systems in multi-turn interaction
gemini-3.1-flash-live-preview (Thinking)
New36.06±4.41
gpt-realtime-1.5
34.73±4.38
gemini-3.1-flash-live-preview
New26.77±4.06
Evaluating spoken dialogue systems in multi-turn interaction
gemini-3-pro-preview (Thinking)
54.65±4.57
gemini-2.5-pro (Thinking)
46.90±4.58
gemini-2.5-flash (Thinking)
40.04±4.50
Evaluating Professional Reasoning in Finance
claude-opus-4-6 (Non-Thinking)
53.28±0.18
Muse Spark
New52.44±0.06
gpt-5
51.32±0.17
Evaluating Professional Reasoning in Legal Practice
Muse Spark
New52.29±0.06
claude-opus-4-6 (Non-Thinking)
52.27±0.66
gpt-5-pro
49.89±0.36
Evaluating AI agents ability to perform real-world, economically valuable remote work
claude-opus-4-6 (CoWork)
New4.17
claude-opus-4-5-20251101-thinking
3.75
Manus_1.6 (Max)
New2.92
Simulating real-world pressure to choose between safe or harmful behavior
o3-2025-04-16
10.50±0.60
claude-sonnet-4-20250514
12.20±0.20
o4-mini-2025-04-16
15.80±0.40
Evaluating how LLMs can dynamically interact with and reason about visual information
gpt-5.4-2026-03-05 (reasoning effort = high)
29.17±0.13
gemini-3.1-pro-preview
28.97±0.91
claude-opus-4-6-thinking
27.52±1.45
Multilingual Native Reasoning Evaluation Benchmark for LLMs
gpt-5-pro-2025-10-06
65.20±1.24
gemini-3.1-pro-preview
64.74±2.88
gpt-5.4-pro-2026-03-05
62.27±2.92
Assessing models across diverse, interdisciplinary challenges
Muse Spark
New75.52±4.05
gemini-3.1-pro-preview
71.37±1.74
gpt-5.4-pro-2026-03-05
69.23±3.05
Frontier Risk Evaluation for National Security and Public Safety
gpt-oss-120b
8.24±1.93
claude-opus-4-5-20251101-thinking
9.63±2.11
claude-sonnet-4-5-20250929-thinking
12.80±2.36
Evaluate model honesty when pressured to lie
claude-opus-4-6 (Non-Thinking)
96.28±0.41
claude-sonnet-4-5-20250929-thinking
96.13±0.57
Claude Sonnet 4 (Thinking)
95.33±2.29
Evaluating model performance on complex, multi-step reasoning tasks
gpt-5.4-pro-2026-03-05
New23.82±2.43
gemini-3.1-pro-preview
New19.76±2.27
gpt-5-pro-2025-10-06
18.75±2.22
Vision-Language Understanding benchmark for multimodal models
Gemini 2.5 Pro Experimental (March 2025)
54.65±1.46
gemini-2.5-pro-preview-06-05
54.63±0.55
gpt-5.4-pro-2026-03-05
New53.89±2.02
Evaluating model performance on common tutoring tasks for high school and AP-level subjects
Muse Spark
New68.55±0.95
gpt-5.4-pro-2026-03-05
56.62±1.02
gemini-2.5-pro-preview-06-05
55.65±1.11
Methodology
We conduct high-complexity evaluations to expose model failures, prevent benchmark saturation, and push model capabilities — while continuously evaluating the latest frontier models.
Humans design complex evaluations and define precise criteria to assess models, while LLMs scale evaluations — ensuring efficiency and alignment with human judgment.
Our leaderboards are built on carefully curated evaluation sets, combining private datasets to prevent overfitting with open-source datasets for broad benchmarking and comparability.
We continuously evaluate the latest releases from every major lab, so the leaderboards stay current with the state of the art — not frozen snapshots.
Evaluate your model
If you'd like to add your model to this leaderboard or a future version, get in touch. To ensure leaderboard integrity, models can only be featured the first time an organization encounters the prompts.
Vahue leaderboard newsletter