AI Model Picker comparison
Gemini 3.7 Flash vs GPT-6 Luna
Both products reached the final shortlist for the same buyer profile — Budget and High Volume — in the AI Model Picker. Your answers favor predictable per-token cost for classification, summarisation, and chat at scale.
Gemini 3.7 Flash
Google budget · Historical checkpoint: see current model pricing and access · Not independently rated
OpenAI
GPT-6 Luna
OpenAI budget · All amounts are USD per MTok; long-context bands apply to the full request, not just excess tokens. Standard <=272K input: input/output $0.10/$0.50, cached input $0.01, 30-minute write $0.125, 12-hour write $0.25; >272K: input/output $0.20/$0.75, cached input $0.02, 30-minute write $0.25, 12-hour write $0.50. Batch <=272K input: input/output $0.05/$0.25, cached input $0.005, 30-minute write $0.0625, 12-hour write $0.125; >272K: input/output $0.10/$0.375, cached input $0.01, 30-minute write $0.125, 12-hour write $0.25. Flex <=272K input: input/output $0.05/$0.25, cached input $0.005, 30-minute write $0.0625, 12-hour write $0.125; >272K: input/output $0.10/$0.375, cached input $0.01, 30-minute write $0.125, 12-hour write $0.25. Fast <=272K input: input/output $0.20/$1.00, cached input $0.02, 30-minute write $0.25, 12-hour write $0.50; >272K: input/output $0.40/$1.50, cached input $0.04, 30-minute write $0.50, 12-hour write $1.00. Regional processing and FedRAMP add 10% where the model and route are eligible; region eligibility is separate from a quoted tariff. Server-side tools and their tokens can add charges. EU data residency supports Standard/Fast/Flex/Batch. Chat Completions function calling requires reasoning_effort none; use Responses for built-in tools. · Not independently rated
| What differs | Gemini 3.7 Flash | GPT-6 Luna |
|---|---|---|
| Vendor | ||
| Vendor | OpenAI | |
| Positioning | ||
| Positioning | Google budget | OpenAI budget |
| Price positioning | ||
| Price positioning | Historical checkpoint: see current model pricing and access | All amounts are USD per MTok; long-context bands apply to the full request, not just excess tokens. Standard <=272K input: input/output $0.10/$0.50, cached input $0.01, 30-minute write $0.125, 12-hour write $0.25; >272K: input/output $0.20/$0.75, cached input $0.02, 30-minute write $0.25, 12-hour write $0.50. Batch <=272K input: input/output $0.05/$0.25, cached input $0.005, 30-minute write $0.0625, 12-hour write $0.125; >272K: input/output $0.10/$0.375, cached input $0.01, 30-minute write $0.125, 12-hour write $0.25. Flex <=272K input: input/output $0.05/$0.25, cached input $0.005, 30-minute write $0.0625, 12-hour write $0.125; >272K: input/output $0.10/$0.375, cached input $0.01, 30-minute write $0.125, 12-hour write $0.25. Fast <=272K input: input/output $0.20/$1.00, cached input $0.02, 30-minute write $0.25, 12-hour write $0.50; >272K: input/output $0.40/$1.50, cached input $0.04, 30-minute write $0.50, 12-hour write $1.00. Regional processing and FedRAMP add 10% where the model and route are eligible; region eligibility is separate from a quoted tariff. Server-side tools and their tokens can add charges. EU data residency supports Standard/Fast/Flex/Batch. Chat Completions function calling requires reasoning_effort none; use Responses for built-in tools. |
| Editorial rating | ||
| Editorial rating | Not independently rated | Not independently rated |
| Best for | ||
| Best for | High-volume Google API work where token efficiency matters more than 3.8 Flash's extra quality. | Classification, summarisation, and chat where unit cost dominates. |
Intelligence comparison
What actually differs
Structured comparison from canonical product intelligence — verification status and recent pricing changes included.
| Criterion | Gemini 3.7 Flash | GPT-6 Luna |
|---|---|---|
| Model | ||
| Context window | 1048576 May be staleSource: Official documentation | 1050000 May be staleSource: Official documentation |
| Pricing | ||
| Input price per 1M tokens | Not verified May be staleSource: Official documentation | $0.10 May be staleSource: Official documentation |
| Output price per 1M tokens | Not verified May be staleSource: Official documentation | $0.50 May be staleSource: Official documentation |
Model
Context window
Gemini 3.7 Flash
1048576
GPT-6 Luna
1050000
Pricing
Input price per 1M tokens
Gemini 3.7 Flash
Not verified
GPT-6 Luna
$0.10
Output price per 1M tokens
Gemini 3.7 Flash
Not verified
GPT-6 Luna
$0.50
Features
What each one leads with
The standout capabilities recorded for each product in the recommendation data.
Gemini 3.7 Flash
- Budget tier in catalog
- Profile comparisons
- Multimodal where verified
GPT-6 Luna
- Lowest OpenAI price in catalog
- 1.05M context on profile
- Cached input $0.01 per 1M
Trade-offs
Pros and cons, side by side
The strengths and the catches the decision tool already weighs for this buyer profile.
Gemini 3.7 Flash
Pros
- Google keeps it for efficiency-first workloads (3.8 Flash can consume more tokens)
- Introductory paid rates through 2026-12-31
Cons
- Hard reasoning still needs Pro-tier escalation
- Introductory price doubles on 2027-01-01
GPT-6 Luna
Pros
- Half GPT-5.6 Luna's $0.20/$1.20
- 20x cheaper than GPT-6 Sol on both input and output
Cons
- Overspecified hard tasks will fail — escalate tier when needed
- DeepSWE 66.6% is OpenAI-reported, not independently run
The verdict
Who should pick which
Assembled from the same recommendation fields the tool scores on — not a universal winner.
Both products are finalists for the same buyer profile, so this is a fit decision rather than a category decision. Gemini 3.7 Flash is aimed at high-volume Google API work where token efficiency matters more than 3.8 Flash's extra quality. GPT-6 Luna is aimed at classification, summarisation, and chat where unit cost dominates. Neither product has an independently verified rating here, so use the fit criteria and current product evidence rather than a synthetic score.
Choose Gemini 3.7 Flash if…
High-volume Google API work where token efficiency matters more than 3.8 Flash's extra quality.
Watch out for
Hard reasoning still needs Pro-tier escalation
Historical checkpoint: see current model pricing and access
Choose GPT-6 Luna if…
Classification, summarisation, and chat where unit cost dominates.
Watch out for
Overspecified hard tasks will fail — escalate tier when needed
All amounts are USD per MTok; long-context bands apply to the full request, not just excess tokens. Standard <=272K input: input/output $0.10/$0.50, cached input $0.01, 30-minute write $0.125, 12-hour write $0.25; >272K: input/output $0.20/$0.75, cached input $0.02, 30-minute write $0.25, 12-hour write $0.50. Batch <=272K input: input/output $0.05/$0.25, cached input $0.005, 30-minute write $0.0625, 12-hour write $0.125; >272K: input/output $0.10/$0.375, cached input $0.01, 30-minute write $0.125, 12-hour write $0.25. Flex <=272K input: input/output $0.05/$0.25, cached input $0.005, 30-minute write $0.0625, 12-hour write $0.125; >272K: input/output $0.10/$0.375, cached input $0.01, 30-minute write $0.125, 12-hour write $0.25. Fast <=272K input: input/output $0.20/$1.00, cached input $0.02, 30-minute write $0.25, 12-hour write $0.50; >272K: input/output $0.40/$1.50, cached input $0.04, 30-minute write $0.50, 12-hour write $1.00. Regional processing and FedRAMP add 10% where the model and route are eligible; region eligibility is separate from a quoted tariff. Server-side tools and their tokens can add charges. EU data residency supports Standard/Fast/Flex/Batch. Chat Completions function calling requires reasoning_effort none; use Responses for built-in tools.
Common questions
Gemini 3.7 Flash vs GPT-6 Luna
Answered from the verified figures on this page rather than general guidance.
Is Gemini 3.7 Flash or GPT-6 Luna cheaper?
Which is rated higher, Gemini 3.7 Flash or GPT-6 Luna?
Should I choose Gemini 3.7 Flash or GPT-6 Luna?
What is the catch with Gemini 3.7 Flash and GPT-6 Luna?
A head-to-head answers one question: of the two finalists for this buyer profile, which fits your situation. If neither “best for” line matches, run the full tool — its other profiles exist for different buyers.