vs
10:0
overall capability wins (1 ties across 11 scored). GPT-5.6 Luna leads overall.
- Reasoning5:0
- Coding2:0
- Arena1:0
- Tool use1:0
- Composite indices2:0
- Pricing3:0
- Speed1:0
- Specs1:0
Verdict
GPT-5.6 Luna leads coding (widest gap: +77.2 on TermBench); GPT-5.6 Luna ranks higher on LMArena (+63 Elo); GPT-5.6 Luna offers ~2.6x larger context (1.1M vs 400K).
- GPT-5.6 Luna leads coding (widest gap: +77.2 on TermBench)
- GPT-5.6 Luna ranks higher on LMArena (+63 Elo)
- GPT-5.6 Luna offers ~2.6x larger context (1.1M vs 400K)
- GPT-5.6 Luna is ~1.5x cheaper on a blended token basis than GPT-5 Mini
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- GPT-5.6 Luna
- $0.50
- GPT-5 Mini
- $0.75
GPT-5.6 Luna is about 1.5x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
OpenAI
OpenAI
Family
GPT-5.x
GPT-5.x
License
Proprietary
Proprietary
Open weights
No
No
Release
Jul 9, 2026
Aug 7, 2025
Knowledge cutoff
-
-
API / provider
OpenAI
OpenAI
Modalities
text, image → text
text, image → text
Specs
Context window
1.1M
400K
Max output
128K
128K
Parameters
—
—
Pricing
Input $/1M
$0.20
$0.25
Output $/1M
$1.20
$2
Blended $/1M (3∶1)
$0.45
$0.69
Speed
tok/s
195
160
TTFT (s)
0.2
0.2
Reasoning
MMLU-Pro
—
—
GPQA Diamond
91.1
82.8
Humanity's Last Exam
39.5
21.5
AIME 2025
—
91.1
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
42.7
25.0
AA-LCR v1.1
83.7
72.3
CritPt
20.6
—
MMMU-Pro
78.6
70.1
IFBench
—
75.4
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
56.2
SWE-bench Pro
62.7
—
SWE-bench Multilingual
—
39.7
LiveCodeBench
—
—
Terminal-Bench 2.1
80.9
3.7
Aider Polyglot
—
71.6
Terminal-Bench 3
14.3
—
BigCodeBench
—
—
SciCode
53.6
39.0
CursorBench
61.1
—
SWE-Rebench
43.6
—
NL2Repo-Bench
—
—
DeepSWE
67.0
—
WebDev Arena
1,519
—
Terminal-Bench 4.0
17.3
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,453
1,390
| Metric | GPT-5.6 Luna | GPT-5 Mini | Delta |
|---|---|---|---|
| Identity | |||
| Organization | OpenAI | OpenAI | — |
| Family | GPT-5.x | GPT-5.x | — |
| License | Proprietary | Proprietary | — |
| Open weights | No | No | — |
| Release | Jul 9, 2026 | Aug 7, 2025 | — |
| Knowledge cutoff | - | - | — |
| API / provider | OpenAI | OpenAI | — |
| Modalities | text, image → text | text, image → text | — |
| Specs | |||
| Context window | 1.1M | 400K | A +650K |
| Max output | 128K | 128K | tie |
| Parameters | — | — | — |
| Pricing | |||
| Input $/1M | $0.20 | $0.25 | A +$0.05 |
| Output $/1M | $1.20 | $2 | A +$0.80 |
| Blended $/1M (3∶1) | $0.45 | $0.69 | A +$0.24 |
| Speed | |||
| tok/s | 195 | 160 | A +35 |
| TTFT (s) | 0.2 | 0.2 | tie |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | 91.1 | 82.8 | A +8.3 pts |
| Humanity's Last Exam | 39.5 | 21.5 | A +18.0 pts |
| AIME 2025 | — | 91.1 | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 42.7 | 25.0 | A +17.7 pts |
| AA-LCR v1.1 | 83.7 | 72.3 | A +11.4 pts |
| CritPt | 20.6 | — | — |
| MMMU-Pro | 78.6 | 70.1 | A +8.5 pts |
| IFBench | — | 75.4 | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | 56.2 | — |
| SWE-bench Pro | 62.7 | — | — |
| SWE-bench Multilingual | — | 39.7 | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | 80.9 | 3.7 | A +77.2 pts |
| Aider Polyglot | — | 71.6 | — |
| Terminal-Bench 3 | 14.3 | — | — |
| BigCodeBench | — | — | — |
| SciCode | 53.6 | 39.0 | A +14.6 pts |
| CursorBench | 61.1 | — | — |
| SWE-Rebench | 43.6 | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | 67.0 | — | — |
| WebDev Arena | 1,519 | — | — |
| Terminal-Bench 4.0 | 17.3 | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,453 | 1,390 | A +63 Elo |
GPT-5.6 Luna
- Context
- 1.1M / 128K out
- Parameters
- —
- Price
- $0.20 / $1.20
- Speed
- 195 tok/s · 0.2s TTFT
- Modalities
- text, image → text
- License
- Proprietary
Fastest, cheapest GPT-5.6 tier for high-volume chat, classification, and light coding; July 30, 2026 API cut dropped Standard rates 80% to $0.20/$1.20.
GPT-5 Mini
- Context
- 400K / 128K out
- Parameters
- —
- Price
- $0.25 / $2
- Speed
- 160 tok/s · 0.2s TTFT
- Modalities
- text, image → text
- License
- Proprietary
Cost-efficient GPT-5 variant for high-throughput applications where flagship quality is not required.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.