Seed 2.0 Mini: benchmarks, price and rank

Its headline results

The benchmarks that count most in the rating, with its place among the models measured the same way, and where each number came from.

• LiveCodeBench v664.1%

  Programming contest problems collected after models were trained, from LeetCode, AtCoder and Codeforces.

  36th of 43. The maker's own number, from ByteDance, via LLM Stats.

• MMMU-Pro71.4%

  College exam questions that need a picture to answer, such as charts, diagrams, maps and chemical structures, with more answer options than the original MMMU and some questions shown only as a screenshot.

  37th of 51. The maker's own number, from ByteDance, via LLM Stats.

• CharXiv Reasoning70.8%

  Questions about real charts from scientific papers that need reasoning across the chart, not just reading off a value.

  39th of 45. The maker's own number, from ByteDance, via LLM Stats.

• MRCR v2 (8 needles)51.4%

  A very long conversation full of similar requests, where the model has to find and repeat particular earlier answers, eight of them at once.

  11th of 24. The maker's own number, from ByteDance, via LLM Stats.

• HealthBench30%

  5,000 realistic health conversations with patients and clinicians, each graded against criteria written by 262 physicians.

  16th of 16. The maker's own number, from ByteDance, via LLM Stats.

• ERQA56.3%

  400 questions about images of the physical world: space, motion, actions, the state of objects and views from several cameras.

  16th of 21. The maker's own number, from ByteDance, via LLM Stats.

• GPQA Diamond79%

  Graduate-level questions in biology, physics and chemistry, written by PhDs so that searching the web does not give the answer away.

  The maker's own number, from ByteDance, via LLM Stats.

• MMLU-Pro83.6%

  Twelve thousand multiple-choice questions across fourteen subjects, from law and medicine to engineering, with ten options each.

  27th of 56. The maker's own number, from ByteDance, via LLM Stats.