Skip to main content
FreeAIList

Compare AI models

Compare documented free access, limits and capabilities for up to four models.

Need a starting point? Use Model Finder
Choose models 4 of 4 selected

Select up to four models. Uncheck one before adding another. Filters narrow the available choices and the comparison.

Filter by capability or access
Available models (14 of 14)

4 visible of 4 selected.

How we compare models

Access and limits

Each column shows documented provider access. On smaller screens, swipe the table or focus it and use the arrow keys to compare columns.

Documented access and limits for selected AI models
CompareGemini 3.8 FlashGoogleGemini 3.7 FlashGoogleGemini 3.6 FlashGoogleGemini 2.5 FlashGoogle
Free access

Gemini API pricing: Free input and output tokens; check AI Studio for current project and model quotas.

Source verified

Gemini API pricing: Free input and output tokens; check AI Studio for current project and model quotas.

Source verified

Gemini API pricing: Free input and output tokens; check AI Studio for current project and model quotas.

Source verified

Gemini API pricing: Free input and output tokens; check AI Studio for current project and model quotas.

Source verified

Reset / contextunknown · 1,048,576 input / 65,536 output tokensunknown · 1,048,576 input / 65,536 output tokensunknown · 1,048,576 input / 65,536 output tokensunknown · 1,048,576 input / 65,536 output tokens
Modalities / usestext, image, audio, video · everyday, reasoning, multimodal, codingtext, image, audio, video · everyday, reasoning, multimodal, codingtext, image, audio, video · everyday, reasoning, multimodal, codingtext, image, audio, video · everyday, reasoning, multimodal
No cardyesyesyesyes
Important limits
  • Google lists a Free Tier; exact rate limits vary by model and project
  • Daily request quotas reset at midnight Pacific time
  • Google lists a Free Tier; exact rate limits vary by model and project
  • Daily request quotas reset at midnight Pacific time
  • Google lists a Free Tier; exact rate limits vary by model and project
  • Daily request quotas reset at midnight Pacific time
  • Free API quotas are model- and project-specific
  • Do not compare provider descriptions with independent benchmark scores without matching versions
Access sources
Last verified

Benchmark observations

Only scores from the same documented comparison group appear together. They are observations, not universal rankings.

OTIS Mock AIME 2024–2025

Metric: accuracy · Unit: percent

Scale: 0–100%

Sources, methods and caveats

  • Gemini 3.8 Flash

    Source verified

    Epoch AI administered the 45-question OTIS Mock AIME 2024–2025 math evaluation. The linked model scorecard reports this accuracy. See https://epoch.ai/benchmarks/otis-mock-aime-2024-2025 for test setup and https://epoch.ai/benchmarks/use-this-data for data license and citation.

    • The model scorecard may show its best result across evaluated settings or providers; this is not a guaranteed score for a specific hosted free route.
    • A 45-question math test does not measure general quality, coding, speed, or multimodal ability.

  • Gemini 3.7 Flash

    Source verified

    Epoch AI administered the 45-question OTIS Mock AIME 2024–2025 math evaluation. The linked model scorecard reports this accuracy. See https://epoch.ai/benchmarks/otis-mock-aime-2024-2025 for test setup and https://epoch.ai/benchmarks/use-this-data for data license and citation.

    • The model scorecard may show its best result across evaluated settings or providers; this is not a guaranteed score for a specific hosted free route.
    • A 45-question math test does not measure general quality, coding, speed, or multimodal ability.

  • Gemini 3.6 Flash

    Source verified

    Epoch AI administered the 45-question OTIS Mock AIME 2024–2025 math evaluation. The linked model scorecard reports this accuracy. See https://epoch.ai/benchmarks/otis-mock-aime-2024-2025 for test setup and https://epoch.ai/benchmarks/use-this-data for data license and citation.

    • The model scorecard may show its best result across evaluated settings or providers; this is not a guaranteed score for a specific hosted free route.
    • A 45-question math test does not measure general quality, coding, speed, or multimodal ability.

Epoch AI reports this 45-question math result for each named model. Scorecards may select the best evaluated setting; hosted routes can differ. Data: Epoch AI, CC BY 4.0 and citation. Test method.