LLM Reference
HLE With ToolsactiveReasoningTool use

HLE With Tools: Humanity's Last Exam — With Tools

Metric: Accuracy (higher is better)Introduced: 2025

Humanity's Last Exam expert-level benchmark measured with tool access. Kept separate from no-tools and GPT-5.5 Pro-compute variants.

Models ranked

7

tracked on this benchmark

Score band

63.9 – 45.1

best → lowest tracked

Snapshot trend

+18.80

Jul 3 → Sep 10 · 1 models

Leaderboard

Tracked models ranked by Accuracy (higher is better).

Compare candidates
#Model variant and provenanceScore
1
DeepSeek V4.1 Flash
Version: HLE with tools Pass@1; max reasoning effortHarness: Not recordedEvaluator: Not recordedObserved: Sep 10, 2026Confidence: Not recordedSource
63.9
2
Claude Opus 4.8
Version: HLE with toolsHarness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource

Notes: Official Anthropic launch benchmark.

57.9
3
Claude Sonnet 5
Version: Humanity's Last Exam with web search, web fetch, programmatic tool calling, and code execution; Claude Opus 4.6 grader; 1M total-token capHarness: Not recordedEvaluator: Not recordedObserved: Jun 30, 2026Confidence: Not recordedSource
57.4
4
Kimi K3
Version: HLE-Full with tools; max reasoning effortHarness: Not recordedEvaluator: Not recordedObserved: Jul 16, 2026Confidence: Not recordedSource

Notes: Model variant: Kimi K3 (max reasoning effort). Benchmark variant: HLE-Full with tools. Harness/evaluator not disclosed by Moonshot. Source posture: Moonshot/Kimi self-reported launch table; confidence: medium because the value is vendor-reported and not independently reproduced. Recommended seed value: 56.

56.0
5
GPT-5.5 Pro
Version: HLE with tools; standard GPT-5.5 weightsHarness: Not recordedEvaluator: Not recordedObserved: Apr 24, 2026Confidence: Not recordedSource

Notes: Standard with-tools variant kept separate from GPT-5.5 Pro-compute HLE.

52.2
6
Agents-A1
Version: Humanity's Last Exam with tools accuracy, vendor-reported on Agents-A1 model cardHarness: Not recordedEvaluator: Not recordedObserved: Jun 26, 2026Confidence: Not recordedSource

Notes: Official Intern Science model card reports Agents-A1 at 47.6 on Humanity's Last Exam with tools. Evaluator/source: Intern Science self-reported model-card/project benchmark table. Public harness details are not fully reproducible from the seed row. Confidence: medium for reported value, low-medium for reproducibility. Recommended seed value: 47.6.

47.6
7
DeepSeek V4 Flash
Version: Humanity's Last Exam with tools pass@1, V4-Flash Think MaxHarness: Not recordedEvaluator: Not recordedObserved: Jul 3, 2026Confidence: Not recordedSource
45.1

How to read this benchmark

This benchmark scores models where higher is better. Use scores for directional filtering and shortlisting, not universal quality ranking; then validate pricing, context window, provider availability, and fit for your workload.

Trust this score when

  • There is a fresh timestamped snapshot (or multiple snapshots) for this benchmark.
  • The model list covers the same version family you can actually deploy today.
  • Top candidates overlap with your required routing and feature requirements.

Be cautious when

  • There is only one benchmark snapshot or the dataset appears stale.
  • The benchmark metric direction is opposite of your decision objective.
  • The score difference between options is narrow and likely within implementation variance.

Related benchmarks

Last reviewed: Jun 17, 2026

Resources