priorsBuilder
Support

Settings

Theme

Explore themes

Mode

Library

Looks borrowed from benchmark sites and products. Themes with a light and a dark version follow your mode.

Bring your own theme

Upload a Priors theme file, or have your agent make one: copy the prompt, tell it the look you want, and upload what it returns. With the Priors skill installed it already knows the format.

Custom themes stay in this browser. They travel inside shared chart links and a page's theme, so others see them as you do.

For the builder, Explore and model pages on this device. Published pages keep the theme their owner picked.

Mode

OpenAI · released 2026-09-03

GPT-6 Astra

Measured on 244 benchmarks across 5 reasoning efforts.

56.9

#3 of 61 models · best at max

Reasoning efforts in this ranking

max56.9#4 of 139
xhigh55.3#6 of 139
high54.8#7 of 139
medium51.5#13 of 139
low43.0#29 of 139

Strengths

  • GPT-6 Astra
  • Gemini 4 Carbon (#1)
  • Field median
bestField medianFrontierCode 2.0-…Terminal-Bench v4…AA-Briefcase v1.1AutomationBench-AATerminal-Bench-Sc…HLE-DiamondCritPt v2 Hard Su…AA-Omniscience v1…AA-Omniscience v1…GDP.pdf
  1. FrontierCode 2.0-AA15%42.4%#2 of 61
  2. Terminal-Bench v4-AA10%59.1%#5 of 61
  3. AA-Briefcase v1.115%1569#15 of 61
  4. AutomationBench-AA5%68.5%#9 of 61
  5. Terminal-Bench-Science-AA15%63.3%#2 of 61
  6. HLE-Diamond10%63.8%#3 of 61
  7. CritPt v2 Hard Subset10%70.5%#1 of 61
  8. AA-Omniscience v1.1 Accuracy10%61.6%#4 of 61
  9. AA-Omniscience v1.1 Non-hallucination5%50.2%#37 of 61
  10. GDP.pdf5%31.0%#3 of 61

Bars show where GPT-6 Astra sits among the 61 ranked models on each benchmark (right is best); the tick is the field median.

Cost and speed

against 61 ranked models

Cost per task

#23 cheapest of 41

$0.27at max

Time per task

#21 quickest of 40

12 minat max

Output speed

#40 fastest of 52

71 tok/sat max

Time to first token

#50 quickest of 52

6.1 minat max

Output tokens per task

#11 fewest of 44

52kat max

Token price

$1.425 per 1M tokens at API prices$0.061 on ChatGPT Pro 10x

97.0% cache read · 2.61% fresh input · 0.05% cache write · 0.38% output

Measured from T3 Code usage in Codex, Oct 27 to Nov 2, high confidence. One app’s users, not everyone’s. $1.470 at the standard 97 / 2.5 / 0.5 mix.

This ranking’s benchmarks

10 of 10 measured

agentic

AutomationBench-AA

241 results
  1. #11max of 24168.5%
  2. #14xhigh of 24167.2%
  3. #18high of 24166.6%
  4. #24medium of 24164.6%
  5. #48low of 24159.1%

Leader: Gemini 4 Carbon · 77.8%

agentic

AA-Briefcase v1.1

247 results
  1. #25max of 2471569
  2. #29xhigh of 2471544
  3. #37high of 2471505
  4. #51medium of 2471462
  5. #93low of 2471256

Leader: Sonnet 5.5 · 1814

reasoning

CritPt v2 Hard Subset

100 results
  1. #1high of 10071.5%
  2. #2medium of 10071.0%
  3. #3max of 10070.5%
  4. #4xhigh of 10070.5%
  5. #17low of 10060.0%

Leader: this model · 71.5%

coding

FrontierCode 2.0-AA

97 results
  1. #3max of 9742.4%
  2. #6xhigh of 9741.4%
  3. #9high of 9740.1%
  4. #16medium of 9736.6%
  5. #34low of 9726.2%

Leader: Opus 5.5 · 46.4%

economics

GDP.pdf

240 results
  1. #2xhigh of 24032.2%
  2. #6max of 24031.0%
  3. #7high of 24031.0%
  4. #10medium of 24030.4%
  5. #11low of 24030.4%

Leader: Muse Wonder Full · 32.6%

reasoning

HLE-Diamond

142 results
  1. #5max of 14263.8%
  2. #7xhigh of 14261.5%
  3. #9high of 14259.3%
  4. #16medium of 14252.7%
  5. #24low of 14245.1%

Leader: Gemini 4 Carbon · 69.1%

knowledge

AA-Omniscience v1.1 Accuracy

111 results
  1. #11max of 11161.6%
  2. #12high of 11161.1%
  3. #14xhigh of 11161.0%
  4. #16medium of 11159.6%
  5. #20low of 11158.6%

Leader: Fable 5.1 · 68.4%

knowledge

AA-Omniscience v1.1 Non-hallucination

111 results
  1. #48high of 11157.5%
  2. #50low of 11155.8%
  3. #51xhigh of 11155.4%
  4. #52medium of 11154.8%
  5. #59max of 11150.2%

Leader: Gemini 4 Carbon · 95.8%

reasoning

Terminal-Bench-Science-AA

115 results
  1. #2max of 11563.3%
  2. #6high of 11559.5%
  3. #10xhigh of 11554.8%
  4. #12medium of 11553.3%
  5. #25low of 11537.1%

Leader: Gemini 4 Carbon · 64.8%

coding

Terminal-Bench v4-AA

238 results
  1. #8xhigh of 23859.6%
  2. #9max of 23859.1%
  3. #17high of 23854.0%
  4. #24medium of 23849.5%
  5. #34low of 23841.9%

Leader: Kimi K3.1 · 67.2%

All 244 benchmarks

show 234 more