GPT-6.2 Astra

openai/gpt-6.2-astra

OpenAI flagship for the hardest engineering & research.

cmd --model gpt-6.2-astra
Intelligence index
not yet scored
Coding index
not yet scored
Input
$8 /M
Output
$40 /M
Cache read
$0.40 /M
Agent-loop cost
$2.68 /M in
Context window
1.05M tokens
Released
November 17, 2026
Modalities
→

vs. the lineup

GPT-6.2 Astra beside its stablemates and nearest rivals. The ◆ marks the best value in each column across every row shown.

pin a rival:
ModelIntelligenceCodingInput $/MOutput $/MBlended $/MContext
GPT-6 Astra52.7◆76.9◆$10$50$201.05M◆
GPT-6.1 Sol51.8—$2$10$41.05M◆
GPT-6 Sol47.6—$2$10$41.05M◆
GPT-6.2 Astra ◆——$8$40$161.05M◆
GPT-6.2 Sol——$2$10$41.05M◆
GPT-6.1 Terra——$1$5$21.05M◆
DeepSeek V4 Pro (latest)pinned3668.8$0.66$1.98$0.991M
DeepSeek V4 Flash (latest)pinned3469.1$0.15◆$0.60◆$0.26◆1M

price bands

GPT-6.2 Astra is not billed at one flat rate — the price changes with how much context a request carries. The header quotes the standard band; here is every band.

BandContextInput $/MOutput $/MCache read $/M
Standard≤ 272K$8.00$40.00$0.40
Long context> 272K$16.00$60.00$0.80

usage calculator

How far a month of credits goes on GPT-6.2 Astra.

GPT-6.2 Astra runs on Max and up — from $100/mo.
Input tokensfresh prompt
800
Output tokensmodel reply
180
Cache read tokensre-read context
50K
cost / request $0.034 · in $8 · out $40 · cache $0.40 per M
fresh input 19%output 21%cache reads 60%
Requests / 30 days
3.0K
$100 credits ÷ $0.034 per request
~595 quick fixes~119 bug fixes~20 feature PRs

what real work costs

Real coding tasks priced end to end on GPT-6.2 Astra, from a quick lookup to a full-repo agent run.

One agent task
$0.89
180K in at 75% cache hit, 12K out
What you pay
$4.66 /M
all-in across every token that task touched
Sticker input
$8 /M
cache reads bill at $0.40 /M instead
TaskTokens in · outGPT-6.2 AstraClaude Opus 5.5Claude Haiku 4.5
Quick lookup / one-liner8K · 1K$0.05$0.02$0.0065
Review a 500-line PR60K · 4K$0.32$0.16$0.04
Fix a bug (agent loop)180K · 12K$0.89$0.45$0.12
Refactor a module320K · 20K$1.61$0.80$0.21
Full-repo agent run900K · 45K$3.66$1.83$0.49

frequently asked

How much does GPT-6.2 Astra cost?
$8/M input and $40/M output, cache reads $0.40/M. In an agent loop most input is cache-read, so the effective input rate is about $2.68/M.
Which plan do I need?
Available on Max and above.
How do I switch to it?
Run cmd --model gpt-6.2-astra, or type /model in a session and pick it. You can switch mid-session without losing context.

Ship code that matches your taste

Command Code is the AI coding agent that continuously learns your taste. Start for $1.

Benchmarks from Artificial Analysis (v4.3)commandcode.ai/models/gpt-6-2-astra