Inkling

Leaderboards

Where Inkling 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 Inkling has a published result on, out of 100%, with its rank among the models that took it.

• SWE-Bench Verified

  #27 / 120

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

  77.6%

  Self-reported

• MMMU-Pro

  #39 / 73

  College-level questions that need both an image and text to answer.

  73.5%

  Self-reported · In the vision score

• CharXiv-R

  #33 / 60

  Reading and reasoning about charts from scientific papers.

  78.1%

  Self-reported

• Terminal-Bench 2.1

  #34 / 46

  Terminal-Bench 2.1 is an updated release of the Terminal-Bench benchmark that tests AI agents' ability to operate a computer via the terminal. It evaluates how well models handle real-world, end-to-end tasks autonomously, including compiling code, training models, setting up servers, system administration, data science workflows, and security tasks.

  63.8%

  Self-reported

• IFBench

  #8 / 44

  Instruction Following Benchmark evaluating model's ability to follow complex instructions

  79.8%

  Self-reported

• MCP Atlas

  #14 / 38

  MCP Atlas is a benchmark for evaluating AI models on scaled tool use capabilities, measuring how well models can coordinate and utilize multiple tools across complex multi-step tasks.

  76.0%

  Self-reported

• AIME 2026

  #2 / 26

  All 30 problems from the 2026 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999. Used as an AI benchmark to evaluate large language models' ability to solve complex mathematical problems requiring multi-step logical deductions and structured symbolic reasoning.

  97.1%

  Self-reported

• Global-MMLU-Lite

  #2 / 16

  A lightweight version of Global MMLU benchmark that evaluates language models across multiple languages while addressing cultural and linguistic biases in multilingual evaluation.

  88.7%

  Self-reported

• GDPval-AA

  #11 / 14

  GDPval-AA evaluates AI agents on economically valuable professional knowledge-work tasks and reports performance as an Elo score.

  41.3%

  Self-reported

• Humanity's Last Exam (no tools, text-only)

  #12 / 14

  Text-only Humanity's Last Exam variant evaluated without tool use.

  29.7%

  Self-reported

• Tau3 Banking

  #4 / 13

  τ³-Bench banking domain evaluates agentic models on multi-turn, tool-using customer-support scenarios in a simulated retail banking environment.

  23.7%

  Self-reported

• Humanity's Last Exam (with tools, text-only)

  #9 / 10

  Text-only Humanity's Last Exam variant evaluated with tool use enabled.

  46.0%

  Self-reported

• MMAU

  #2 / 5

  A massive multi-task audio understanding and reasoning benchmark comprising 10,000 carefully curated audio clips paired with human-annotated natural language questions spanning speech, environmental sounds, and music. Requires expert-level knowledge and complex reasoning across 27 distinct skills.

  77.2%

  Self-reported

• SimpleQA Verified

  #2 / 5

  SimpleQA Verified is a curated, reliability-focused subset of SimpleQA that addresses label noise and redundancy in the original benchmark, measuring short-form parametric factual accuracy of large language models on fact-seeking questions with single, indisputable answers.

  43.9%

  Self-reported

• AA-Omniscience Index

  #3 / 4

  AA-Omniscience Index is Artificial Analysis's knowledge-reliability metric. It rewards correct answers, penalizes hallucinations, and does not penalize abstention. Scores range from -100 to 100, where 0 means as many correct as incorrect answers.

  1.1%

  Self-reported

• VoiceBench Avg

  #2 / 4

  VoiceBench is the first benchmark designed to provide a multi-faceted evaluation of LLM-based voice assistants, evaluating capabilities including general knowledge, instruction-following, reasoning, and safety using both synthetic and real spoken instruction data with diverse speaker characteristics and environmental conditions.

  91.4%

  Self-reported