readable AI benchmarks (simplified)

readable AI benchmarks

OpenAI

GPT-6.2 Astra (High)

reasoning · closed weights · Nov 17, 2026

GPT-6.2 Astra (High) has a Quality Score of 66.7, ranking 11th among 283 scored models. Per 1M tokens, pricing is $8.00 input, $40.00 output, $0.400 cached input, and $10.00 cache write. It has a 1M token context window. Its Reliability Score is 73.4, ranking 15th among 288 scored models, above average (total average: 48.2). Its Value Score is 38.0, high compared with the total Value average of 29.3.

Quality Score66.7coverage 100.0%: AA Intelligence and AA-Omniscience outcomes available
Value Score38.0
Factual reliability73.4
Cache Discount95%

Model specification

Reasoning
reasoning
Input modalities
none
Output modalities
none
Context window
1M tokens
Weights
closed weights

Token prices USD per 1M tokens

Input
$8.00
Output
$40.00
Cached Input
$0.400
Cache write
$10.00
Cost per task
$2.07

Capability

Intelligence
60.1
Omniscience (native net)
46.8
Correct
66.3%
Blended price
$5.88

Answer outcomes

Fully graded outcomes (Correct + Incorrect): 85.9%

Correct 66.3%Incorrect 19.6%Partial / not attempted 14.1%

Artificial Analysis benchmarks aa-v5

Current Intelligence Index components

AA-Briefcase v1.1
1570.4 Elo
AutomationBench-AA
76.0%
private 657-task split from dataset v1.0.6
GDP.pdf
28.6%
FrontierCode v2.0
47.1%
AA Stirrup · 132 tasks, 3 attempts per task
Terminal-Bench v4.0
60.1%
AA mini-swe-agent standardized runs · 66 tasks, pass@1 averaged over 3 repeats per task
AA-Omniscience v1.1
46.8 native net
HLE-Diamond
65.9%
CritPt v2 Hard
71.5%
AA Stirrup · hard subset, 40 of 70 problems
Terminal-Bench-Science v0.1.0
75.2%

Similar models

Support me! Patreon