GPT-5.4 Pro vs GPT-6 Astra

Change models
IndexAll categories, out of 100.
73.3 ±8.7
78.0 ±1.9 best
GapIndex points below the top.
−4.7 · no clear gap
Top
CoverageIndex weight that has results.
60%
100% best
SpeedOutput tokens per second.
21/s
34/s best
EloLMArena rating and rank.
—
1440
ReleasedMonth of release.
2026-03
2026-09
Input / 1MUS dollars per 1M input tokens.
$30.00
$10.00 best
Output / 1MUS dollars per 1M output tokens.
$180.00
$50.00 best
Batch output / 1MBatch price per 1M output tokens.
$90.00
$25.00 best
1K chats2K tokens in, 500 out, each.
$150.00
$45.00 best
1K documents20K tokens in, 1K out, each.
$780.00
$250.00 best
1K agent tasks200K tokens in, 20K out, each.
$9,600.00
$3,000.00 best
ContextMaximum tokens in one request.
1.05M
1.05M
AgenticMulti-step tasks with tools.
70.5
74.6 best
CodingCode writing and repair.
—
76.9
ReasoningLogic problems and puzzles.
80.4
83.8 best
MultimodalTasks with images and text.
—
71.2
KnowledgeFacts and expert knowledge.
67.2
78.5 best
MultilingualTasks in many languages.
—
73.6
InstructionTasks with strict prompt rules.
—
79.5
MathMath problems.
75.3
78.0 best

Benchmarks

11 shared
Only GPT-5.4 Pro: 7Only GPT-6 Astra: 127
GPT-5.4 Pro
GPT-6 Astra
94.6%8 / 217
95.8% best3 / 217
30.0%10 / 214
31.7% best4 / 214
83.3%19 / 131
95.0% best1 / 131
94.5%21 / 126
97.5% best5 / 126
92.2%12 / 122
97.5% best5 / 122
98.3% best8 / 116
97.8%14 / 116
46.3%40 / 91
75.6% best4 / 91
Chess PuzzlesReasoning
59.0%10 / 90
74.0% best3 / 90
82.5%17 / 90
93.7% best2 / 90
58.5%20 / 72
97.6% best3 / 72
BrowseCompAgentic
89.3%11 / 63
91.5% best6 / 63