← Model list

GPT-5-Codex

openai·openai/gpt-5-codex·GA·Closed·gpt-codex series

Coding-optimized GPT model for repository edits, reviews, and agentic software work

At a glance
Reference price$1.25 / $10.00 per 1M
Cache read $0.125 · Abacus
Lowest paid$1.07 / $8.50
OpenCode Zen Gateway · 1.2× spread
Context400,000
⚠ Providers report 256,000–400,000; the table below is authoritative
Output limit128,000
Capabilities
ReasoningTool useStructured outputTemperatureAttachments
⚠ Providers report capability flags inconsistently
Modalities
TextImagePDF
Knowledge cutoff2024-09-30
Released / updated2025-09-15 / 2025-09-15

Quality & performanceArtificial Analysis · Intelligence Index v4.1 · rep. tier high

Intelligence37
Value13
Coding
Agentic
Output speed
TTFT
Value formulaIQ 37 ÷ min blended $2.928 = 13

Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.

Available at 11 providers11 with public prices

ProviderTierInputOutputCache readCache writeContextOutput limitStatus
OpenCode ZenGateway$1.07$8.50$0.107400,000128,000
PoeGateway$1.10$9.00400,000128,000
Jiekou.AIGateway$1.13$9.00400,000128,000
OrcaRouterGateway$1.25$10.00$0.125400,000128,000
NanoGPTGateway$1.25$10.00$0.125256,00032,768
ImpossiblGateway$1.25$10.00$0.125400,000128,000
ZenMuxGateway$1.25$10.00$0.12400,00064,000
Azure Cognitive ServicesFirst-party$1.25$10.00$0.13400,000128,000
AzureFirst-party$1.25$10.00$0.13400,000128,000
Vercel AI GatewayCloud$1.25$10.00$0.13400,000128,000
AbacusGateway$1.25$10.00$0.125400,000128,000

Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.

Your usage cost

1OpenCode Zen$523.44
2ZenMux$614.40
3OrcaRouter$615.00
4NanoGPT$615.00
5Impossibl$615.00
6Abacus$615.00
The cheapest paid channel is the only channel.

Benchmark3 items

NameConditionsScoreMetricSource
Artificial Analysis Coding Index38.9indexSource ↗
SciCode40.9percent correctSource ↗
Terminal-Bench Hard37.9success rateSource ↗

The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.

Reasoning control

effort = loweffort = mediumeffort = high

4 / 11 providers expose no reasoning control (reasoning_options: []).

Related models

Price historyone sample accumulated per data sync

Input listOutput listMin blended
$10.00$02026-08-052026-08-132026-08-05 · Input list $1.252026-08-06 · Input list $1.252026-08-07 · Input list $1.252026-08-08 · Input list $1.252026-08-09 · Input list $1.252026-08-10 · Input list $1.252026-08-11 · Input list $1.252026-08-12 · Input list $1.252026-08-13 · Input list $1.252026-08-05 · Output list $10.002026-08-06 · Output list $10.002026-08-07 · Output list $10.002026-08-08 · Output list $10.002026-08-09 · Output list $10.002026-08-10 · Output list $10.002026-08-11 · Output list $10.002026-08-12 · Output list $10.002026-08-13 · Output list $10.002026-08-05 · Min blended $2.932026-08-06 · Min blended $2.932026-08-07 · Min blended $2.932026-08-08 · Min blended $2.932026-08-09 · Min blended $2.932026-08-10 · Min blended $2.932026-08-11 · Min blended $2.932026-08-12 · Min blended $2.932026-08-13 · Min blended $2.93