GPT-6 Astra Benchmarks, Specifications & Availability

67 · raw 67 index

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

Artificial Analysis Coding Agent Index v1.4 as listed by OpenAI

59.3% · raw 59.3 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

33.3% · raw 33.3 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-27

As reported by NaiveAI.

34.2% · raw 34.2 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology | Differs from current headline 59.3 (first-party); kept with provenance, non-headline.

98.5% · raw 98.5 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

ARC-AGI-1

95% · raw 95 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

ARC-AGI-2

99.9% · raw 99.9 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

41.4% · raw 41.4 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

41.4% · raw 41.4 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology | Same value as current headline; kept as corroboration, non-headline.

74.1% · raw 74.1 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

74.1% · raw 74.1 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology | Same value as current headline; kept as corroboration, non-headline.

100% · raw 100 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

39% · raw 39 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

Internal contamination-aware ExploitBench Jun–Aug 2026

42.4% · raw 42.4 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

64.5% · raw 64.5 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

FrontierCode 1.1 Extended; footnote 8 on OpenAI page

53.3% · raw 53.3 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

FrontierCode 1.1 Main; footnote 8

53.3% · raw 53.3 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-22

As reported by Anthropic Opus 5.5 announce | Demoted 2026-09-25: duplicate of headline from OpenAI GPT-6 Astra announce (openai.com/index/gpt-6-astra); vendor official preferred over Anthropic peer comparison table (rule a); same value.

97.6% · raw 97.6 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

FrontierMath Tier 4 (v2); prose also cites 98% saturation

37.1% · raw 37.1 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

GeneBench Pro

96% · raw 96 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

64.7% · raw 64.7 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-22

length-adjusted (raw 68.2, mean length 3185 chars); Sep 22 system-card correction supersedes earlier launch-post 63.4

70.3% · raw 70.3 %

Alternative · max effort · Peer vendor

Source/record date: 2026-09-28

As reported by Anthropic (Sonnet 5.5 card); length-adjusted with Anthropic grader. Kept non-headline under Astra own-card 64.7 (different harness)

57.2% · raw 57.2 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

57.2% · raw 57.2 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-22

As reported by Anthropic Opus 5.5 announce | Demoted 2026-09-25: duplicate of headline from OpenAI GPT-6 Astra announce (openai.com/index/gpt-6-astra); vendor official preferred over Anthropic peer comparison table (rule a); same value.

72.6% · raw 72.6 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

v2026.08.08 offline set, partial score; ~40 min/task

72.6% · raw 72.6 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology | Same value as current headline; kept as corroboration, non-headline.

92.7% · raw 92.7 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

ScreenSpot-Pro no tools

85.4% · raw 85.4 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

88% · raw 88 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

SRE-Bench reverse engineering

57.9% · raw 57.9 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

57.9% · raw 57.9 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-22

As reported by Anthropic Opus 5.5 announce (OpenAI figure) | Demoted 2026-09-25: duplicate of headline from OpenAI GPT-6 Astra announce (openai.com/index/gpt-6-astra); vendor official preferred over Anthropic peer comparison table (rule a); same value.

58.2% · raw 58.2 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology | Differs from current headline 57.9 (first-party); kept with provenance, non-headline.

64.6% · raw 64.6 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

64.6% · raw 64.6 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-22

As reported by Anthropic Opus 5.5 announce (OpenAI figure) | Demoted 2026-09-25: duplicate of headline from OpenAI GPT-6 Astra announce (openai.com/index/gpt-6-astra); vendor official preferred over Anthropic peer comparison table (rule a); same value.

68.1% · raw 68.1 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology | Differs from current headline 64.6 (first-party); kept with provenance, non-headline.

36.6% · raw 36.6 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-22

length-adjusted (raw 34.2, mean length 1697 chars); Astra system card HealthBench table (Sep 22 corrected values)

1542 · raw 1542 Elo

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-22

As reported by Anthropic Opus 5.5 announce

41.4% · raw 41.4 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-22

As reported by Anthropic Opus 5.5 announce (Zapier public)

61 · raw 61 index

Headline · unknown effort · Peer vendor

Source/record date: Not recorded

from chart/figure as reported on Ling-3.0-flash-VL HF card AA Index v4.1.1 chart (peer spillover); GPT-6 Astra (max)

65.5% · raw 65.5 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-28

As reported by Anthropic (Sonnet 5.5 card); Proximal harness

71% · raw 71 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-28

As reported by Anthropic (Sonnet 5.5 card); without tools, per Surge AI public reporting

71% · raw 71 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology | Same value as current headline; kept as corroboration, non-headline.

60% · raw 60 %

Headline · max effort · Peer vendor

Source/record date: 2026-09-28

As reported by Anthropic (Sonnet 5.5 card); length-adjusted Sep-2026 figure with Anthropic grader and harness, differs from Astra own-card setup

63.1% · raw 63.1 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology

53.5% · raw 53.5 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology

5.4% · raw 5.4 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology

65.5% · raw 65.5 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology

89.6% · raw 89.6 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology

44.3% · raw 44.3 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology

85.4% · raw 85.4 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology

72% · raw 72 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology

98.7 f1 · raw 98.7 f1

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology

71.8 f1 · raw 71.8 f1

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology

87.5% · raw 87.5 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology

68% · raw 68 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology

8.5% · raw 8.5 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon chart; peer harness notes in Google eval methodology

53 · raw 53 index

Headline · max effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard;

52 · raw 52 index

Alternative · xhigh effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard;

51 · raw 51 index

Alternative · high effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard;

50 · raw 50 index

Alternative · medium effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard;

46 · raw 46 index

Alternative · low effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard;

$3.26 · raw 3.26 USD

Headline · max effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard;

$2.31 · raw 2.31 USD

Alternative · xhigh effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard;

$1.73 · raw 1.73 USD

Alternative · high effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard;

$1.54 · raw 1.54 USD

Alternative · medium effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard;

$0.82 · raw 0.82 USD

Alternative · low effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard;

54 tok/s · raw 54 tok/s

Headline · max effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard; Median output tokens/s; leaderboard rounds to whole tokens.

49 tok/s · raw 49 tok/s

Alternative · xhigh effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard; Median output tokens/s; leaderboard rounds to whole tokens.

45 tok/s · raw 45 tok/s

Alternative · high effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard; Median output tokens/s; leaderboard rounds to whole tokens.

45 tok/s · raw 45 tok/s

Alternative · medium effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard; Median output tokens/s; leaderboard rounds to whole tokens.

46 tok/s · raw 46 tok/s

Alternative · low effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard; Median output tokens/s; leaderboard rounds to whole tokens.

62 · raw 62 index

Headline · unknown effort · Independent evaluator

Source/record date: 2026-09-22

Codex harness; ties Fable 5.1

63.1% · raw 63.13 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-03

Refreshed from current Vals leaderboard. Cost/test $18.46.

45 fixes · raw 45 fixes

Headline · max effort · Independent evaluator

Source/record date: 2026-10-03

Harness: Codex CLI; effort max; 3 runs; evaluation 2026-09-14. Best documented score for this effort in Oct 1 README. Headline: best documented model run.

43 fixes · raw 43 fixes

Alternative · xhigh effort · Independent evaluator

Source/record date: 2026-10-03

Harness: Codex CLI; effort xhigh; 1 runs; evaluation 2026-09-04. Best documented score for this effort in Oct 1 README.

35 fixes · raw 35 fixes

Alternative · high effort · Independent evaluator

Source/record date: 2026-10-03

Harness: Codex CLI; effort high; 1 runs; evaluation 2026-09-04. Best documented score for this effort in Oct 1 README.

34 fixes · raw 34 fixes

Alternative · medium effort · Independent evaluator

Source/record date: 2026-10-03

Harness: Codex CLI; effort medium; 1 runs; evaluation 2026-09-05. Best documented score for this effort in Oct 1 README.

27 fixes · raw 27 fixes

Alternative · low effort · Independent evaluator

Source/record date: 2026-10-03

Harness: Codex CLI; effort low; 1 runs; evaluation 2026-09-05. Best documented score for this effort in Oct 1 README.

89.6% · raw 89.59 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-03

Refreshed from current Vals leaderboard. Harness: OpenHands. Cost/test $38.51.

97.5% · raw 97.5 %

Alternative · max effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration.

98.5% · raw 98.5 %

Headline · xhigh effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration. Headline: best verified value; source matrix effort order breaks ties.

98.5% · raw 98.5 %

Alternative · high effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration.

97.5% · raw 97.5 %

Alternative · medium effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration.

96.5% · raw 96.5 %

Alternative · low effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration.

86% · raw 86 %

Alternative · none effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration.

95% · raw 95 %

Headline · max effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration. Headline: best verified value; source matrix effort order breaks ties.

93.3% · raw 93.3 %

Alternative · xhigh effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration.

92.1% · raw 92.1 %

Alternative · high effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration.

92.1% · raw 92.1 %

Alternative · medium effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration.

85.4% · raw 85.4 %

Alternative · low effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration.

59.6% · raw 59.6 %

Alternative · none effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration.

62.7% · raw 62.71 %

Headline · max effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Harness: Standard (notes carry). Headline: best verified value; source matrix effort order breaks ties.

59.3% · raw 59.34 %

Alternative · xhigh effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Harness: Standard (notes carry).

54.8% · raw 54.82 %

Alternative · high effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Harness: Standard (notes carry).

38.6% · raw 38.59 %

Alternative · medium effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Harness: Standard (notes carry).

17.5% · raw 17.45 %

Alternative · low effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Harness: Standard (notes carry).

35.2% · raw 35.18 %

Alternative · none effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Harness: Standard (notes carry).

98.6% · raw 98.55 %

Alternative · max effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Harness: Provider Adapter.

98.4% · raw 98.44 %

Alternative · xhigh effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Harness: Provider Adapter.

100% · raw 99.95 %

Headline · high effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Harness: Provider Adapter. Headline: best verified value; source matrix effort order breaks ties.

98.4% · raw 98.44 %

Alternative · medium effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Harness: Provider Adapter.

98% · raw 98.03 %

Alternative · low effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Harness: Provider Adapter.

96.7% · raw 96.72 %

Alternative · none effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Harness: Provider Adapter.

44 · raw 44 index

Headline · high effort · Independent evaluator

Source/record date: 2026-09-23

AA-Omniscience Index; GPT-6 Astra (high) as cited on evaluation page

1542 · raw 1542 Elo

Headline · max effort · Independent evaluator

Source/record date: 2026-09-23

GDPval-AA v2.1 Elo; GPT-6 Astra (max)

$15014.70 · raw 15014.7 $

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-03

Vending-Bench 2 net gain = final_value in the page public vb2 data module minus $500 starting balance. Full 66-model source checked.

49.7% · raw 49.7 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-03

Blueprint-Bench 2 connectivity similarity; published fractional score multiplied by 100.

42.2% · raw 42.22 %

Headline · xhigh effort · Independent evaluator

Source/record date: 2026-10-02

WeirdML variant GPT-6 Astra (xhigh); harness codex_cli 0.154.0; values from prepared data JSON; raw 0.422202; official 80/20 aggregate (area 500k-50M tokens + final best)

54.1% · raw 54.05 %

Headline · xhigh effort · Independent evaluator

Source/record date: 2026-10-02

WeirdML variant GPT-6 Astra (xhigh); harness codex_cli 0.154.0; values from prepared data JSON; raw 0.540453; mean final best effective score

$27.85 · raw 27.85 USD

Headline · xhigh effort · Independent evaluator

Source/record date: 2026-10-02

WeirdML variant GPT-6 Astra (xhigh); harness codex_cli 0.154.0; values from prepared data JSON; mean API cost per run, same task weighting as scores