vibecodingnav
中/EN
Models/gpt-6-astra

GPT-6 Astra

First on EQ-Bench Creative Writing v3 (Elo 2176.9). Second on WebDev at 1773. Tied at 58.2% on Terminal-Bench. Fifth on sudo L7 at 31.1%.

OpenAI · Proprietary · Context 1.1M · $10 / $50 per million tokens · Fetched 2026-10-10

EQ-Bench Creative Writing v3
2176.9
#1 / 146
LMArena WebDev
1773
#2 / 153 · ±10
LMArena Image-to-WebDev
1726
#3 / 65 · ±12

Benchmark scores

Boards older than 90 days are not counted as recommendations
BenchmarkSettingScoreRankCostAs of
EQ-Bench Creative Writing v3—2176.9#1—2026-10-29
LMArena WebDevmax1773±10#2—2026-10-27
LMArena Image-to-WebDevmax1726±12#3—2026-10-19
Android Benchcodex28.013.3–42.0#4$375.72026-10-26
Terminal-Bench 4.0max · Codex58.2±2.8#3$3.3k2026-10-06
The Slop Index—58#29—2026-10-28
sudo L7Max reasoning31.1#5—2026-10-28
Next.js Agent Evalshigh · Codex90.0#8$0.9692026-10-30
Convex Evals3 runs89.88#1$6.052026-10-28
FrontierCodemax · codex42.8#3$6.682026-10-27
DeepSWE v1.1xhigh · mini-swe-agent74.171.2–77.0#1$4.432026-09-22
SWE-bench Verified—This board has stopped updating——2026-02-26
Aider Polyglot—This board has stopped updating——2025-11-20

Use it or not

Why
  • Still first on Creative Writing Elo; after Slop inversion the docs primary board is led by Inkling Small / the Haiku band.
  • 37 points behind first on WebDev.
When not
  • Output is $50 per million tokens. Nearby ranks cost half of that, or less.

Appears in these scenarios

Collected from scenario boards
Build a Web frontend#1DeepSWE 74.1%Build an Android app#1DeepSWE 74.1%Write CLI, scripts, and DevOps#1DeepSWE 74.1%Write docs & copy#1Creative Writing 2176.9

Peers

ModelCreative WritingPrice $/M
GPT-6 AstraCurrent2176.9$10 / $50Kimi K3.1 Pro2172.6$2 / $10Claude Fable 5.12155.3$10 / $50