GLM-4.5

Leaderboards

Where GLM-4.5 sits among the top 20 in each category.

By task

Scored from votes on each kind of prompt alone, on the same 1,000-point scale.

Test results

Every test GLM-4.5 has a published result on, out of 100%, with its rank among the models that took it.

• GPQA

  #107 / 256

  Graduate-level science questions written to be hard to look up.

  79.1%

  Self-reported

• MMLU-Pro

  #28 / 142

  Broad knowledge and reasoning across 14 subjects, ten answer choices each.

  84.6%

  Self-reported

• SWE-Bench Verified

  #86 / 120

  Fixing real bugs in real GitHub projects, checked by the projects' own tests.

  64.2%

  Self-reported

• Humanity's Last Exam

  #87 / 108

  Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions

  14.4%

  Self-reported · In the text score· In the vision score

• LiveCodeBench

  #21 / 76

  LiveCodeBench is a holistic and contamination-free evaluation benchmark for large language models for code. It continuously collects new problems from programming contests (LeetCode, AtCoder, CodeForces) and evaluates four different scenarios: code generation, self-repair, code execution, and test output prediction. Problems are annotated with release dates to enable evaluation on unseen problems released after a model's training cutoff.

  72.9%

  Self-reported

• BrowseComp

  #66 / 68

  Finding hard-to-locate facts by browsing the web.

  26.4%

  Self-reported · In the agents score

• AIME 2024

  #8 / 54

  American Invitational Mathematics Examination 2024, consisting of 30 challenging mathematical reasoning problems from AIME I and AIME II competitions. Each problem requires an integer answer between 0-999 and tests advanced mathematical reasoning across algebra, geometry, combinatorics, and number theory. Used as a benchmark for evaluating mathematical reasoning capabilities in large language models at Olympiad-level difficulty.

  91.0%

  Self-reported

• MATH-500

  #3 / 34

  MATH-500 is a subset of the MATH dataset containing 500 challenging competition mathematics problems from AMC 10, AMC 12, AIME, and other mathematics competitions. Each problem includes full step-by-step solutions and spans multiple difficulty levels across seven mathematical subjects including Prealgebra, Algebra, Number Theory, Counting and Probability, Geometry, Intermediate Algebra, and Precalculus.

  98.2%

  Self-reported

• SciCode

  #17 / 30

  SciCode is a research coding benchmark curated by scientists that challenges language models to code solutions for scientific problems. It contains 338 subproblems decomposed from 80 challenging main problems across 16 natural science sub-fields including mathematics, physics, chemistry, biology, and materials science. Problems require knowledge recall, reasoning, and code synthesis skills.

  41.7%

  Self-reported

• TAU-bench Retail

  #6 / 25

  A benchmark for evaluating tool-agent-user interaction in retail environments. Tests language agents' ability to handle dynamic conversations with users while using domain-specific API tools and following policy guidelines. Evaluates agents on tasks like order cancellations, address changes, and order status checks through multi-turn conversations.

  79.7%

  Self-reported

• Terminal-Bench

  #12 / 25

  Terminal-Bench is a benchmark for testing AI agents in real terminal environments. It evaluates how well agents can handle real-world, end-to-end tasks autonomously, including compiling code, training models, setting up servers, system administration, security tasks, data science workflows, and cybersecurity vulnerabilities. The benchmark consists of a dataset of ~100 hand-crafted, human-verified tasks and an execution harness that connects language models to a terminal sandbox.

  37.5%

  Self-reported

• TAU-bench Airline

  #4 / 23

  Part of τ-bench (TAU-bench), a benchmark for Tool-Agent-User interaction in real-world domains. The airline domain evaluates language agents' ability to interact with users through dynamic conversations while following domain-specific rules and using API tools. Agents must handle airline-related tasks and policies reliably.

  60.4%

  Self-reported

• BFCL-v3

  #1 / 20

  Berkeley Function Calling Leaderboard v3 (BFCL-v3) is an advanced benchmark that evaluates large language models' function calling capabilities through multi-turn and multi-step interactions. It introduces extended conversational exchanges where models must retain contextual information across turns and execute multiple internal function calls for complex user requests. The benchmark includes 1000 test cases across domains like vehicle control, trading bots, travel booking, and file system management, using state-based evaluation to verify both system state changes and execution path correctness.

  77.8%

  Self-reported

• AA-Index

  #1 / 4

  No official academic documentation found for this benchmark. Extensive research through ArXiv, IEEE/ACL/NeurIPS papers, and university research sites yielded no peer-reviewed sources for an 'aa-index' benchmark. This entry requires verification from official academic sources.

  67.7%

  Self-reported