GPT-6 LunavsClaude Sonnet 5
Across 3 shared benchmarks, GPT-6 Luna leads overall: GPT-6 Luna wins 2, Claude Sonnet 5 wins 1, with 0 ties and an average score difference of -16.21.
GPT-6 Luna
OpenAI · 2026-09-22 · Reasoning model
Claude Sonnet 5
Anthropic · 2026-06-30 · Multimodal model
GPT-6 Luna2 wins(67%)(33%)1 winClaude Sonnet 5
Benchmark scores
Grouped by capability, sorted by largest gap within each. 3 shared benchmarks.
AI Agent - Tool Usage
GPT-6 Luna 1/1| Benchmark | GPT-6 Luna | Claude Sonnet 5 | Diff |
|---|---|---|---|
| Terminal-Bench 4.0 | 1348 / 95Max (With Tools) | 12.4252 / 95Max (With Tools) | +0.58 |
Multimodal Understanding
GPT-6 Luna 1/1| Benchmark | GPT-6 Luna | Claude Sonnet 5 | Diff |
|---|---|---|---|
| GDP.pdf | 2040 / 122Max (No Tools) | 13.2068 / 122Max (No Tools) | +6.80 |
Productivity Knowledge
Claude Sonnet 5 1/1| Benchmark | GPT-6 Luna | Claude Sonnet 5 | Diff |
|---|---|---|---|
| AA-Briefcase | 1,29938 / 88Max (With Tools) | 1,35532 / 88Max (With Tools) | -56 |
Specs
| Field | GPT-6 Luna | Claude Sonnet 5 |
|---|---|---|
| Publisher | OpenAI | Anthropic |
| Release date | 2026-09-22 | 2026-06-30 |
| Model type | Reasoning model | Multimodal model |
| Architecture | Dense | Dense |
| Parameters | Not available | Not available |
| Context length | 1.05M | 1M |
| Max output | 128K | 128K |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | GPT-6 Luna | Claude Sonnet 5 |
|---|---|---|
| Text input | $0.1 / 1M tokens | $2 / 1M tokens |
| Text output | $0.5 / 1M tokens | $10 / 1M tokens |
| Cache read | $0.01 / 1M tokens | $0.2 / 1M tokens |
| Cache write | $0.125 / 1M tokens | $2.5 / 1M tokens |
Summary
- GPT-6 Lunaleads in:AI Agent - Tool Usage (1/1), Multimodal Understanding (1/1)
- Claude Sonnet 5leads in:Productivity Knowledge (1/1)
On average across the 3 shared benchmarks, Claude Sonnet 5 scores 16.21 higher.
Largest single-benchmark gap: AA-Briefcase — GPT-6 Luna 1,299 vs Claude Sonnet 5 1,355 (-56).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.