Patchwire

OpenAI · Released Feb 24, 2026

GPT-5.3-Codex

GPT-5.3-Codex is OpenAI’s most advanced agentic coding model, combining the frontier software engineering performance of GPT-5.2-Codex with the broader reasoning and professional knowledge capabilities of GPT-5.2. It achieves…

ReasoningVision
  • Top 15%Terminal-Bench
  • Top 31%METR time horizon
  • Top 35%Vending-Bench 2
Compare in showdown → OpenRouter ↗ All OpenAI models
Released
Feb 24 2026
Context window
400K 128K out
Input / 1M tokens
$1.75
Output / 1M tokens
$14.00

Where to run it

2 providers compared

Cheapest: Azure, OpenAI at $1.75 / $14.00 per 1M input / output tokens — 50% less than the priciest option.

ProviderInputOutputCache readContextMax outputUptimevs. best
Azure $1.75 $14.00 $0.17400K128K100%24h Cheapest
OpenAI $1.75 $14.00 $0.17400K128K100%24h Cheapest
OpenAIfast $3.50 $28.00 $0.35400K128K— +100%

Value

Price vs. performance

7 models score higher for the same price or lesson SWE-bench Verified, e.g. Gemini 3.5 Flash.
57.6%66.2%74.8%83.5% $0.10$1.00$10.00$100.00 DeepSeek V4 Pro 0813 (DeepSeek) — 77.6% · $0.66/1M in GLM 5.2 (Z.ai) — 78.7% · $0.62/1M in Qwen3.7 Max (Qwen) — 77.3% · $1.48/1M in Gemini 3.5 Flash (Google) — 79.3% · $1.50/1M in DeepSeek V4 Pro 0423 (DeepSeek) — 77.6% · $0.95/1M in Kimi K2.6 (Moonshot AI) — 76.6% · $0.65/1M in Claude Opus 4.7 (Anthropic) — 83.5% · $5.00/1M in GLM 5.1 (Z.ai) — 74.2% · $0.97/1M in Qwen3.6 Plus (Qwen) — 57.9% · $0.33/1M in Gemini 3.1 Pro Preview Custom Tools (Google) — 75.6% · $2.00/1M in Claude Sonnet 4.6 (Anthropic) — 75.2% · $3.00/1M in GLM 5 (Z.ai) — 72.1% · $0.60/1M in Claude Opus 4.6 (Anthropic) — 78.7% · $5.00/1M in Kimi K2.5 (Moonshot AI) — 73.8% · $0.50/1M in Gemini 3 Flash Preview (Google) — 75.4% · $0.50/1M in Claude Opus 4.5 (Anthropic) — 76.6% · $5.00/1M in Claude Sonnet 4.5 (Anthropic) — 71.3% · $3.00/1M in Claude Opus 4.1 (Anthropic) — 73.4% · $15.00/1M in Gemini 2.5 Pro (Google) — 57.6% · $1.25/1M in Gemini 2.5 Pro Preview 06-05 (Google) — 57.6% · $1.25/1M in GPT-5.4 (OpenAI) — 76.9% · $2.50/1M in GPT-5.2 (OpenAI) — 73.8% · $1.75/1M in GPT-5.1 (OpenAI) — 68.0% · $1.25/1M in GPT-5 (OpenAI) — 73.6% · $1.25/1M in GPT-5 Mini (OpenAI) — 64.7% · $0.25/1M in GPT-5.3-Codex (OpenAI) — 74.8% · $1.75/1M in GPT-5.3-Codex

GPT-5.3-Codex Other OpenAI Other labs Value frontier · SWE-bench Verified vs. input price per 1M tokens (log scale) · Drag to pan, pinch or Ctrl + scroll to zoom, double-click to zoom in.

Benchmarks

5 benchmark scores

BenchmarkScoreRankLeader
SWE-bench VerifiedSoftware engineering · Real GitHub issues fixed in Python repos 74.8% #14of 26 Claude Opus 4.783.5%
Terminal-BenchSoftware engineering · Tasks completed with the programs in a real terminal (best agent) 78.4% #6of 40 GPT-5.584.7%
METR time horizonSoftware engineering · Length of software task solved more often than not 5.8 h #4of 13 Claude Opus 4.612 h
Vending-Bench 2Agents & computer use · Run a simulated vending business for a year (final balance) $5,940 #21of 61 GPT-6 Astra$15,515
WebDev ArenaDesign & visual · People vote on which model builds the better website (Elo) 1370 #71of 88 Claude Opus 5.51820

Specs

Details

Model ID
openai/gpt-5.3-codex
Developer
OpenAI
Release date
Feb 24, 2026 ·
Context window
400,000 tokens
Max output
128,000 tokens
API price
$1.75 / $14.00 per 1M input / output tokens
Providers
2
Input types
text, image, file
Reasoning
Yes
Tool calling
Yes
Open weights
No
Free tier
No

Release date, prices and providers from OpenRouter's catalog; buying directly from a provider can cost a little differently. Uptime is the share of OpenRouter requests each provider served over the last day. Benchmark scores are the best result per benchmark from Epoch AI, licensed CC BY 4.0. Rank counts every model tracked here with a score.