Back to all models
G

Gemini Ultra 1.0

GoogleReleased 2023-12-06
Official Page

Google's original most capable model, designed to be a general-purpose AI that outperforms experts on 30 of 32 benchmarks.

Overall Score
78%
Cost Efficiency
100
Context Window
128K

Strengths

  • ✓Strong math and reasoning
  • ✓Good knowledge scores
  • ✓Multimodal capabilities

Weaknesses

  • ✗Outdated compared to newer models
  • ✗Weak coding performance
  • ✗No longer available to general users
Architecture
Transformer (MoE)
Parameters
Unknown
Input
$0/M
Output
$0/M

Benchmark Performance

BenchmarkDescriptionScoreCategoryBar
MMLUMassive Multitask Language Understanding90%knowledge
HumanEvalCode generation accuracy74.4%coding
MATHMathematical reasoning94.4%math
GSM8KGrade school math94.4%math
GPQAGraduate-level science Q&A59.4%reasoning
HellaSwagCommon sense reasoning87.2%reasoning
SWE-benchSoftware engineering problems26.9%coding
ARC-CAI2 Reasoning Challenge96%reasoning
AI
ModelBench

The open benchmark platform for comparing the latest AI models and their capabilities.

Platform

  • Compare Models
  • All Models
  • Methodology

AI Labs

  • OpenAI
  • Anthropic
  • Google DeepMind
  • Meta AI

Resources

  • Documentation
  • Benchmark Guide
  • API Reference
Benchmarks are sourced from official model cards and independent evaluations. Data is for informational purposes only.
© 2026 ModelBench. Built for the AI community.