Responses and Chat Completions are supported; the Decisions API was added in beta on October 6, 2026. Evaluate that endpoint separately and recheck prices and token tiers.
The provider announcement describes intended uses, not an independent result in your application. Check accepted inputs, thinking controls and current pricing on the selected Gemini endpoint.
Use a fixed dataset and a separate difficult subset. Luna launched September 22, 2026, and Gemini 3.8 Flash reached general availability September 2. Match the application behavior while keeping native endpoint differences documented.
Treat endpoint choice as part of the experiment
Luna can be evaluated through Responses or Chat Completions; Decisions became a separate beta option October 6. Gemini uses its own request and thinking controls. Save the chosen endpoint and configuration with every result.
Include repair in the outcome
Record invalid structures, wrong values, missing fields and retries separately. Then calculate spend per correct accepted result. A fast schema-compliant answer can still create expensive manual work.
Three criteria for this comparison
Labeled decision and extraction accuracy
Actual endpoint and thinking configuration
Accepted-result cost under realistic load
Run a fair comparison
Use the same representative inputs and a written answer rubric. Pin the model version and record the endpoint, thinking configuration, quantization when known, context length, output allowance and concurrency. Preserve native request differences in an adapter rather than silently dropping fields.
Measure full request time, error rate and reported usage. For tools, validate arguments and score the result after execution. For images, preserve resolution and preprocessing. A valid response format alone does not establish a correct answer.
Compare cost per accepted result
Calculate input cost plus all billed output, including reasoning where the provider counts it. Add retries and discarded answers. Divide the total by the number of results that pass your quality criteria. Keep API usage cost separate from deployment and engineering cost.
Latest recorded measurements aligned by exact artifact, benchmark version, source and methodology. These are source results, not tests of Flonno's deployment.
Artificial Analysis observations are refreshed by the integrated Radar. Differences in artifact, benchmark version or method prevent a score comparison. Read the evidence methodology →
YOUR WORKLOAD
Compare cost per accepted task.
Enter results from your own evaluation. A task should count once it meets your written quality criteria. Include failed attempts and retries in total spend. These inputs stay in this browser and are not a Flonno benchmark.
GPT-6 Luna
Enter all four values. Attempted tasks must be at least one; accepted tasks cannot exceed attempted tasks.
Gemini 3.8 Flash
Enter all four values. Attempted tasks must be at least one; accepted tasks cannot exceed attempted tasks.
Turn evidence into a deployment decision
A source benchmark can inform your shortlist, while your evaluation shows whether a model finishes the work at an acceptable cost. Keep the chosen endpoint, processing terms and output limits in the decision record. Inclusion in the Radar does not mean the model is served by Flonno.
