GPT-5.6 Luna vs GPT-6 Astra

Change models
IndexAll categories, out of 100.
65.5 ±2.1
78.0 ±1.9 best
GapIndex points below the top.
−12.5
Top
CoverageIndex weight that has results.
100%
100%
SpeedOutput tokens per second.
51/s best
34/s
EloLMArena rating and rank.
1431
1440 best
ReleasedMonth of release.
2026-07
2026-09
Input / 1MUS dollars per 1M input tokens.
$0.20 best
$10.00
Output / 1MUS dollars per 1M output tokens.
$1.20 best
$50.00
Batch output / 1MBatch price per 1M output tokens.
$0.60 best
$25.00
1K chats2K tokens in, 500 out, each.
$1.00 best
$45.00
1K documents20K tokens in, 1K out, each.
$5.20 best
$250.00
1K agent tasks200K tokens in, 20K out, each.
$64.00 best
$3,000.00
ContextMaximum tokens in one request.
1.05M
1.05M
AgenticMulti-step tasks with tools.
64.7
74.6 best
CodingCode writing and repair.
67.4
76.9 best
ReasoningLogic problems and puzzles.
66.3
83.8 best
MultimodalTasks with images and text.
65.8
71.2 best
KnowledgeFacts and expert knowledge.
62.1
78.5 best
MultilingualTasks in many languages.
53.1
73.6 best
InstructionTasks with strict prompt rules.
48.2
79.5 best
MathMath problems.
68.4
78.0 best

Benchmarks

78 shared
Only GPT-5.6 Luna: 24Only GPT-6 Astra: 60
GPT-5.6 Luna
GPT-6 Astra
39.5%49 / 222
54.7% best8 / 222
83.7% best10 / 217
80.7%37 / 217
91.6%31 / 217
95.8% best3 / 217
20.6%29 / 214
31.7% best4 / 214
EuroEval PolishMultilingual
54.2%50 / 209
67.4% best2 / 209
98.3%28 / 200
100.0% best1 / 200
91.1%27 / 199
96.1% best1 / 199
EuroEval SwedishMultilingual
66.9%15 / 195
77.6% best1 / 195
EuroEval GermanMultilingual
52.6%27 / 191
68.0% best1 / 191
60.7%27 / 182
74.7% best2 / 182
EuroEval FrenchMultilingual
68.9%10 / 178
76.3% best2 / 178
EuroEval SpanishMultilingual
57.9%19 / 176
70.0% best1 / 176
EuroEval DutchMultilingual
59.3%28 / 175
74.7% best1 / 175
EuroEval ItalianMultilingual
60.0%30 / 174
77.8% best2 / 174
42.7%42 / 164
61.6% best6 / 164
92.3%18 / 141
96.0% best1 / 141
48.0%30 / 133
53.7% best25 / 133
59.5%43 / 131
95.0% best1 / 131
88.0%42 / 126
97.5% best5 / 126
77.1%39 / 123
89.6% best8 / 123