Head-to-head comparison

GPT-6 Luna vs. GPT-5.6 Terra

Compare published benchmark results from matching versions and cohorts, API pricing, and technical specifications. Any documented test setup differences remain visible.

shared result series: 16
distinct benchmarks: 16
latest retrieval: 2026-09-23

First modelSecond model

Specifications

General specifications

Release date, availability, context window, architecture, and published parameter counts at a glance. Undisclosed details remain marked as unavailable.

Specifications for GPT-6 Luna and GPT-5.6 Terra
Attribute	GPT-6 Luna	GPT-5.6 Terra
Provider	OpenAI	OpenAI
Released	Sep 22, 2026	Jul 9, 2026
Availability	Active	Active
Model type	Proprietary	Proprietary
Parameters	Not published	Not published
Architecture	Not published	Not published
Context window	1,050,000 tokens	1,050,000 tokens
Knowledge cutoff	May 18, 2026	February 16, 2026
Technical sources	Model, Context	Model, Context, Knowledge cutoff

Pricing

API pricing compared

The table compares official standard API rates per 1M tokens, including published cache pricing. Batch, fast, priority, regional surcharges, tool calls, and cloud platform pricing are excluded.

API pricing for GPT-6 Luna and GPT-5.6 Terra
Price type	GPT-6 Luna	GPT-5.6 Terra
API input	$0.1 ≤272K / $0.2 >272K	$2 ≤272K / $4 >272K
API output	$0.5 ≤272K / $0.75 >272K	$12 ≤272K / $18 >272K
Cache read	$0.01 ≤272K / $0.02 >272K	$0.2 ≤272K / $0.4 >272K
Cache write (5 min.)	$0.13 ≤272K / $0.25 >272K	$2.5 ≤272K / $5 >272K
Cache write (1 hr.)	Not listed	Not listed
Price verified	09/22/2026 Source	09/22/2026 Source

GPT-6 LunaGPT-5.6 Terra

API inputUSD per 1M tokens, base tier

GPT-6 Luna$0.1

GPT-5.6 Terra$2

API outputUSD per 1M tokens, base tier

GPT-6 Luna$0.5

GPT-5.6 Terra$12

Cache readUSD per 1M tokens, base tier

GPT-6 Luna$0.01

GPT-5.6 Terra$0.2

Key to the marks:better valuebetter, but not conclusivetie

Pricing by provider

Direct developer pricing and provider endpoint tariffs routed through OpenRouter are listed separately. Prices are in USD per 1M tokens. Context-dependent direct tiers plus regional, Flex, Priority, and other OpenRouter tariffs remain individually identifiable.

Provider pricing for GPT-6 Luna and GPT-5.6 Terra
Model	Provider	Purchase route and tariff	Input per 1M tokens	Output per 1M tokens	Source
GPT-6 Luna	OpenAI	Direct from developerUp to 272,000 context tokens	$0.1	Lowest listed price: $0.5	Source09/22/2026
	OpenAI	Direct from developerAbove 272,000 context tokens	$0.2	$0.75	Source09/22/2026
	Amazon Bedrock	Via OpenRouteramazon-bedrock/us-east-1	$0.11	$0.55	Source09/23/2026
	Azure	Via OpenRouterazure/eu	$0.11	$0.55	Source09/23/2026
	Azure	Via OpenRouterazure/us	$0.11	$0.55	Source09/23/2026
	OpenAI	Via OpenRouteropenai	Lowest listed price: $0.1	Lowest listed price: $0.5	Source09/23/2026
	OpenAI	Via OpenRouteropenai/fast	$0.2	$1	Source09/23/2026
GPT-5.6 Terra	OpenAI	Direct from developerUp to 272,000 context tokens	$2	$12	Source09/22/2026
	OpenAI	Direct from developerAbove 272,000 context tokens	$4	$18	Source09/22/2026
	Amazon Bedrock	Via OpenRouteramazon-bedrock/us-east-1	$2.2	$13.2	Source09/23/2026
	Azure	Via OpenRouterazure	$2	$12	Source09/23/2026
	Azure	Via OpenRouterazure/eu	$2.2	$13.2	Source09/23/2026
	Azure	Via OpenRouterazure/us	$2.2	$13.2	Source09/23/2026
	OpenAI	Via OpenRouteropenai	$2	$12	Source09/23/2026
	OpenAI	Via OpenRouteropenai/fast	$4	$24	Source09/23/2026
	OpenAI	Via OpenRouteropenai/flex	Lowest listed price: $1	Lowest listed price: $6	Source09/23/2026

Performance

Shared benchmarks

Only published results with the same benchmark version, task, metric, and comparison cohort are paired. Different reasoning levels, agents, harnesses, or output limits appear directly in the table.

Capability profile from matched benchmarks

Each axis averages directly matched benchmark families. A value of 100 means the stronger value within this pair, not a universal quality score.

GPT-6 Luna

GPT-5.6 Terra

Coding: 88.1 / 97.8 (4 benchmark families)

Mathematics: 86.5 / 100 (1 benchmark family)

Finance: 94 / 98.9 (3 benchmark families)

Legal: 86.6 / 64.3 (2 benchmark families)

Agentic tasks: 94.2 / 100 (1 benchmark family)

Productivity: 92.4 / 100 (1 benchmark family)

Health: 100 / 98.1 (2 benchmark families)

Education: 100 / 97.7 (1 benchmark family)

16 matched rows from 16 benchmark families were considered. The radar shows 8 of 9 comparable categories.

GPT-6 LunaGPT-5.6 Terra

Code MigrationOverall. Higher is better. No documented setup difference

GPT-6 Luna42.554 %

GPT-5.6 Terra47.804 %

EMBOverall. Higher is better. No documented setup difference

GPT-6 Luna68.515 %

GPT-5.6 Terra66.205 %

Finance Agent (v2)Overall. Higher is better. No documented setup difference

GPT-6 Luna49.873 %

GPT-5.6 Terra54.436 %

Harvey's Legal Agent BenchmarkOverall · Task fully resolved. Higher is better. No documented setup difference

GPT-6 Luna2.917 %

GPT-5.6 Terra0.833 %

IOIOverall. Higher is better. No documented setup difference

GPT-6 Luna55.556 %

GPT-5.6 Terra87.611 %

Legal Research BenchOverall · All-pass. Higher is better. No documented setup difference

GPT-6 Luna30.288 %

GPT-5.6 Terra41.346 %

MedCodeOverall. Higher is better. No documented setup difference

GPT-6 Luna44.685 %

GPT-5.6 Terra43.414 %

MedScribeOverall. Higher is better. No documented setup difference

GPT-6 Luna83.71 %

GPT-5.6 Terra82.869 %

ProgramBenchOverall. Higher is better. No documented setup difference

GPT-6 Luna0.5 %

GPT-5.6 Terra0.5 %

ProofBenchOverall. Higher is better. No documented setup difference

GPT-6 Luna64 %

GPT-5.6 Terra74 %

Public Benefits Bench v1.1Overall. Higher is better. No documented setup difference

GPT-6 Luna57.645 %

GPT-5.6 Terra62.382 %

SAGEOverall. Higher is better. No documented setup difference

GPT-6 Luna48.091 %

GPT-5.6 Terra47.003 %

Key to the marks:better valuebetter, but not conclusivetie

Benchmark scores for GPT-6 Luna and GPT-5.6 Terra
Benchmark and task	GPT-6 Luna	GPT-5.6 Terra	Source
Code MigrationOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	42.554 %	Winner47.804 %	Vals AI Source details Data as of: 09/22/2026
EMBOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner68.515 %	66.205 %	Vals AI Source details Data as of: 09/22/2026
Finance Agent (v2)Overall Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is better	49.873 %	Winner54.436 %	Vals AI Source details Data as of: 09/22/2026
Harvey's Legal Agent BenchmarkOverall · Task fully resolved Test details No documented difference in the test setupVersion: 1Metric: task resolution rateScoring: Higher is better	Winner2.917 %	0.833 %	Vals AI Source details Data as of: 09/22/2026
IOIOverall Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is better	55.556 %	Winner87.611 %	Vals AI Source details Data as of: 09/22/2026
Legal Research BenchOverall · All-pass Test details No documented difference in the test setupVersion: 1Metric: all-pass rateScoring: Higher is better	30.288 %	Winner41.346 %	Vals AI Source details Data as of: 09/22/2026
MedCodeOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner44.685 %	43.414 %	Vals AI Source details Data as of: 09/22/2026
MedScribeOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner83.71 %	82.869 %	Vals AI Source details Data as of: 09/22/2026
ProgramBenchOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Tie0.5 %	Tie0.5 %	Vals AI Source details Data as of: 09/22/2026
ProofBenchOverall Test details No documented difference in the test setupVersion: 1.1Metric: accuracyScoring: Higher is better	64 %	Winner74 %	Vals AI Source details Data as of: 09/22/2026
Public Benefits Bench v1.1Overall Test details No documented difference in the test setupVersion: 1.1Metric: accuracyScoring: Higher is better	57.645 %	Winner62.382 %	Vals AI Source details Data as of: 09/22/2026
SAGEOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner48.091 %	47.003 %	Vals AI Source details Data as of: 09/22/2026
Tax Agent BenchOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	58.864 %	Winner65.199 %	Vals AI Source details Data as of: 09/22/2026
Terminal-Bench 2.1Overall Test details No documented difference in the test setupVersion: 2.1Metric: accuracyScoring: Higher is better	73.034 %	Winner77.528 %	Vals AI Source details Data as of: 09/22/2026
Vals IndexOverall Test details No documented difference in the test setupVersion: 2Metric: weighted index scoreScoring: Higher is better	58.45 %	Winner59.588 %	Vals AI Source details Data as of: 09/22/2026
Vibe Code Bench v1.1Overall Test details No documented difference in the test setupVersion: 1.1Metric: accuracyScoring: Higher is betterSettings: Harness: OpenHands	Winner81.649 %	74.594 %	Vals AI Source details Data as of: 09/22/2026
Key to the marks:better valuebetter, but not conclusivetie

Our model tests

Our tests

Compare how the models respond to the same prompt. Each test shows the first attempt, with no subsequent fixes to the generated code. These results do not contribute to an overall score.

Pelican on a bicycle

A community classic for free-form SVG drawing.

GPT-6 Luna

Not tested yet

GPT-5.6 Terra

Reasoning (requested): high

First attempt

Run date: 7 Sept 2026
API provider: openrouter
Requested API model: openai/gpt-5.6-terra
API model reported in response: openai/gpt-5.6-terra
Reasoning setting: high
Pinned endpoint: OpenAI | openai/gpt-5.6-terra-20260709
Temperature: Provider default
Token limit including reasoning: 8,192 tokens
Test protocol: community-visual-tests-v1
Technical checks: Passed, not a quality rating
Response time: 64.6 s

Visually reviewed. This individual test does not establish overall model quality.

Prompt and test conditions

Generate an SVG of a pelican riding a bicycle

Token limit including reasoning: 8,192 tokens

Task origin (Simon Willison)

How to read this comparison

Winning one benchmark is not an overall verdict

Your workload, budget, and required context length matter most. A coding benchmark says little about visual understanding or agent performance.

The standard price table uses direct developer rates. The provider comparison labels OpenRouter endpoints, regions, and special tariffs separately.

Every result links to its measurement source and retrieval date. Labels distinguish vendor reports, official benchmark leaderboards, and independent evaluations. Disputed or archived results are excluded.

More matchups

Compare GPT-6 Luna and GPT-5.6 Terra with other leading models using the same data and benchmark logic.

View the complete LLM comparison