GPT-6 Astra (max): benchmark scores, price and ranking

GPT-6 Astra (max) from OpenAI ranks #2 of 741 LLMs on the LLMs Tiger aggregate benchmark score (63.8), based on 23 benchmark results captured up to 2026-10-05.

Organisation
OpenAI
Type
Frontier model
Open weights
No
Released
2026-09-03 (source: epoch)
Cheapest API price
$10.00 in / $50.00 out per million tokens (via azure, observed 2026-10-05T21:47Z)
Aggregate rank
#2 of 741 (score 63.8)

Relative to other models, GPT-6 Astra (max) is strongest on GPQA Diamond (#1 of 274) and weakest on LMArena Text Elo (#16 of 28).

GPT-6 Astra (max) benchmark results

GPT-6 Astra (max) benchmark scores, ranked against all models measured on each benchmark
BenchmarkScoreRankSourceCaptured
GPQA Diamond95.8%#1 of 274Epoch AI Benchmarking Hub2026-10-05
Chess Puzzles72.0%#1 of 124Epoch AI Benchmarking Hub2026-10-05
ARC-AGI-295.0%#1 of 89ARC Prize official leaderboard2026-10-05
ALE-Bench2,951#1 of 78Epoch AI Benchmarking Hub2026-10-05
FrontierMath T1-393.7%#1 of 69Epoch AI Benchmarking Hub2026-10-05
SimpleQA Verified75.6%#1 of 66Epoch AI Benchmarking Hub2026-10-05
Mystery Games84.0%#1 of 57Epoch AI Benchmarking Hub2026-10-05
WeirdML93.3%#2 of 112Epoch AI Benchmarking Hub2026-10-05
OTIS Mock AIME100.0%#5 of 188Epoch AI Benchmarking Hub2026-10-05
CritPt31.7%#3 of 111Epoch AI Benchmarking Hub2026-10-05
BALROG68.3%#1 of 34Epoch AI Benchmarking Hub2026-10-05
WebDev Arena1,800#2 of 63Epoch AI Benchmarking Hub2026-10-05
FrontierMath T497.6%#2 of 52Epoch AI Benchmarking Hub2026-10-05
EBR-Bench76.2%#1 of 19Epoch AI Benchmarking Hub2026-10-05
FrontierSWE65.5%#1 of 16Epoch AI Benchmarking Hub2026-10-05
LMArena Code Elo1,788#2 of 29LMArena2026-10-05
Furniture Assembly80.0%#2 of 25Epoch AI Benchmarking Hub2026-10-05
FrontierCode53.3%#2 of 15Epoch AI Benchmarking Hub2026-10-05
SciCode56.5%#14 of 100Epoch AI Benchmarking Hub2026-10-05
DeepSWE73.2%#3 of 19Epoch AI Benchmarking Hub2026-10-05
LMArena Document Elo1,468#3 of 18LMArena2026-10-05
LMArena Vision Elo1,284#10 of 22LMArena2026-10-05
LMArena Text Elo1,477#16 of 28LMArena2026-10-05

GPT-6 Astra (max) is also listed at 7 other reasoning-effort settings; this page shows the highest-effort row, which is the one the leaderboard keeps by default.

Models ranked nearby

Compare it on the live leaderboard → · How the aggregate score works