[LEADERBOARDS]
testing the limits of AI.
Benchmarks for frontier, agentic, and safety capabilities
Including benchmarks on agentic coding, frontier reasoning, and safety alignment.
From leading AI labs including OpenAI, Anthropic, Google, Meta, and open-source contributors.
Humanity's Sixth Sense
HSS, in partnership with Elorian, is a benchmark for intuitive visual reasoning across diverse image and video inputs.
Gemini-4-Argon (high)
NEW55.40±4.20
GPT-6-Astra (max)
NEW53.60±4.10
GPT-6.1-Sol (max)
NEW46.60±3.80
DistressBench
We introduce DistressBench, designed to measure the quality of crisis support a model delivers.
muse-spark-1.2
88.39±0.93
muse-spark-1.3
84.88±0.93
muse-glimmer-30b
80.57±0.96
Humanity's Last Exam (Diamond)
Challenging LLMs at the frontier of human knowledge
60.60±3.00
55.00±3.10
gemini-4-argon
NEW53.20±3.07
DrugDiscoveryBench
DrugDiscoveryBench: 82 expert-curated tasks evaluating how reliably frontier coding agents perform the computational work of early-stage drug discovery.
GPT 6 Astra (mini-SWE-agent) max
NEW68.70±2.50
GPT 6 Astra (Codex) max
NEW65.40±1.90
Gemini 4 Argon (mini-swe-agent) high
NEW65.40±2.90
SWE Atlas - Refactoring
Evaluating an agent's ability to restructure code while preserving behavior.
GPT 6 Astra (Codex) xHigh
NEW59.05±6.43
Fable-5.1 (Claude Code) xHigh
NEW56.67±6.52
Fable-5 (Claude Code) xHigh
54.76±6.76
SWE Atlas - Test Writing
Evaluating an agent’s ability to write production-grade tests
Fable-5.1 (Claude Code) xHigh*
NEW67.04±5.33
Opus 5 (Claude Code) xHigh
62.22±5.58
Fable-5 (Claude Code) xHigh*
55.60±5.80
SWE Atlas - Codebase QnA
Evaluating deep code comprehension and reasoning
Opus 5 (Claude Code) xHigh
63.17±5.01
Fable-5.1 (Claude Code) xHigh*
NEW59.95±4.85
GPT 6 Astra (Codex) xHigh*
NEW59.14±4.88
HiL-Bench (Human-in-Loop Benchmark)
Evaluates whether agents recognize information gaps and ask targeted clarifying questions.
Muse Wonder 1.0
NEW79.67±4.17
MuseSpark 1.3
66.89±5.12
GPT 6 Astra
65.32±5.19
MCP Atlas
Evaluating real-world tool use through the Model Context Protocol (MCP)
Muse Spark 1.1
88.10±1.95
Fable 5.1
NEW87.20±2.05
claude-opus-5 (xhigh)
85.80±2.10
SWE-Bench Pro V2 Full
Evaluating long-horizon software engineering tasks in public open source repositories
Opus 5 (Claude Code) xhigh
99.40±0.40
Gemini 4 Argon (mini-swe-agent) high
NEW99.22±0.45
Fable 5.1 (Claude Code) high
99.10±0.50
SWE-Bench Pro V2 HARD
Evaluating long-horizon software engineering tasks in public open source repositories
Opus 5 (Claude Code) xhigh
98.00
Fable 5.1 (Claude Code) high
92.20
Gemini 4 Argon (mini-swe-agent) high
NEW92.16
AudioMultiChallenge
Evaluating spoken dialogue systems in multi-turn interaction
gemini-3.8-flash (high)
60.40±4.49
salesforce-launa-omni-1
NEW59.69±4.13
gemini-4-argon (high)
NEW58.41±4.53
AudioMultiChallenge - Audio Output
Evaluating spoken dialogue systems in multi-turn interaction
gpt-realtime-2 (xHigh)
48.45±4.59
tml-interaction-small
43.36±4.55
gpt-realtime-2
37.61±4.45
AudioMultiChallenge - Text Output
Evaluating spoken dialogue systems in multi-turn interaction
gemini-3.8-flash (high)
NEW60.40±4.49
salesforce-launa-omni-1
NEW59.69±4.13
gemini-4-argon (high)
NEW58.41±4.53
Professional Reasoning Benchmark - Finance
Evaluating Professional Reasoning in Finance
Muse Spark 1.3
NEW59.54±1.66
Muse Wonder 1.0
NEW58.67±1.69
Muse Spark 1.1
55.01±0.14
Professional Reasoning Benchmark - Legal
Evaluating Professional Reasoning in Legal Practice
Muse Spark 1.3
NEW61.56±1.72
Muse Spark Open
NEW59.76±0.88
Muse Wonder 1.0
NEW58.40±1.76
Remote Labor Index (RLI)
Evaluating AI agents ability to perform real-world, economically valuable remote work
Opus 5.5
NEW21.25
GPT 6 Astra
20.83
Fable 5.1
17.92
MultiChallenge
Assessing models across diverse, interdisciplinary challenges
GPT 6.1 Sol
NEW81.95±4.16
Gemini 4 Argon
NEW78.70±5.30
Muse Spark
75.52±4.05
Fortress
Frontier Risk Evaluation for National Security and Public Safety
5.78±1.26
8.24±1.93
9.48±1.34
Frontier AI Model Evaluations & Benchmarks
We conduct high-complexity evaluations to expose model failures, prevent benchmark saturation, and push model capabilities--while continuously evaluating the latest frontier models.
Scaling with Human Expertise
Humans design complex evaluations and define precise criteria to assess models, while LLMs scale evaluations--ensuring efficiency and alignment with human judgment.
Robust Datasets for Reliable AI Benchmarks
Our leaderboards are built on carefully curated evaluation sets, combining private datasets to prevent overfitting and open-source datasets for broad benchmarking and comparability.
[EVALUATE YOUR MODEL]
If you'd like to add your model to this leaderboard or a future version, please contact [email protected]. To ensure leaderboard integrity, we require that models can only be featured the FIRST TIME when an organization encounters the prompts.