GPT-5.3-Codex-Spark delivers ultra-fast real-time AI coding at 1,000 tokens per second

AI Coding News

GPT-5.3-Codex-Spark delivers ultra-fast real-time AI coding at 1,000 tokens per second
CerebrasCodex-SparkLow Latency
  • 📰 IntEngineering
  • ⏱ Reading Time:
  • 154 sec. here
  • 11 min. at publisher
  • 📊 Quality Score:
  • News: 89%
  • Publisher: 63%

OpenAI launches GPT-5.3-Codex-Spark, a real-time coding AI delivering 1,000 tokens per second.

OpenAI has launched GPT-5.3- Codex-Spark , its first AI model built specifically for real-time coding, capable of generating more than 1,000 tokens per second while handling real-world software engineering tasks.

The model is a smaller version of GPT-5.3-Codex and is being released as a research preview for ChatGPT Pro users. It is optimized for ultra-low latency performance and runs on specialized hardware developed in partnership with Cerebras.Unlike larger frontier models designed for long-running autonomous tasks, Codex-Spark focuses on instant interaction. Developers can make targeted edits, reshape logic, refine interfaces, and see changes immediately. The model is designed for collaborative coding sessions where speed matters as much as intelligence.At launch, Codex-Spark supports a 128k context window and is text-only. It operates under separate rate limits during the preview phase, and usage does not count toward standard limits. However, users may experience temporary queuing during periods of high demand.Speed meets intelligenceCodex-Spark is tuned for interactive workflows. It makes minimal, focused edits by default and does not automatically run tests unless instructed. This lightweight working style allows developers to interrupt or redirect the model mid-task and iterate quickly.On software engineering benchmarks such as SWE-Bench Pro and Terminal-Bench 2.0, Codex-Spark demonstrates strong accuracy while completing tasks in a fraction of the time compared to GPT-5.3-Codex. The speed advantage comes from both model optimization and infrastructure upgrades.OpenAI implemented end-to-end latency improvements across its serving pipeline. These changes reduced client-server roundtrip overhead by 80 percent, per-token overhead by 30 percent, and time-to-first-token by 50 percent.A persistent WebSocket connection is enabled by default for Codex-Spark and will soon extend to other models.Powered by wafer-scale AICodex-Spark runs on Cerebras’ Wafer Scale Engine 3, a purpose-built AI accelerator optimized for high-speed inference. The partnership adds a low-latency serving tier to OpenAI’s production stack.“What excites us most about GPT-5.3-Codex-Spark is partnering with OpenAI and the developer community to discover what fast inference makes possible—new interaction patterns, new use cases, and a fundamentally different model experience. This preview is just the beginning,” said Sean Lie, CTO and Co-Founder of Cerebras.GPUs remain central to OpenAI’s broader training and inference systems, delivering cost-effective performance at scale. Cerebras hardware complements that setup by focusing on extremely low latency workflows. The two systems can also be combined for single workloads to balance speed and efficiency.Codex-Spark includes the same safety training as OpenAI’s mainline models, including cyber-related safeguards.According to the company’s evaluation process, the model does not meet thresholds for high-risk capability in cybersecurity or biology.The release marks the first step toward a dual-mode Codex system that blends real-time collaboration with longer-horizon reasoning. Future updates are expected to expand capabilities, including larger models, longer context windows, and multimodal inputs.

We have summarized this news so that you can read it quickly. If you are interested in the news, you can read the full text here. Read more:

IntEngineering /  🏆 287. in US

Cerebras Codex-Spark Low Latency Openai Real-Time Inference SWE-Bench Wafer Scale Engine 3

 

United States Latest News, United States Headlines

Similar News:You can also read news stories similar to this one that we have collected from other news sources.

Cardi B and Stefon Diggs Spark Breakup Rumors After Instagram UnfollowCardi B and Stefon Diggs Spark Breakup Rumors After Instagram UnfollowSpeculation arises about the relationship status of Cardi B and New England Patriots player Stefon Diggs following their apparent unfollowing on Instagram after the Super Bowl. The couple, who recently welcomed a baby boy, has fueled rumors of a breakup after fans noticed the social media change and observed Cardi B's attendance at NFL games and her recent activities.
Read more »

Cardi B and Stefon Diggs Spark Breakup Rumors After Super Bowl Loss and Social Media UnfollowingCardi B and Stefon Diggs Spark Breakup Rumors After Super Bowl Loss and Social Media UnfollowingSpeculation arises about the relationship between Cardi B and Stefon Diggs after they unfollowed each other on Instagram following the Super Bowl. Fans are keen to see if this is the end of the couple's relationship.
Read more »

Spark pushes DeFi stablecoin liquidity into institutional crypto lendingSpark pushes DeFi stablecoin liquidity into institutional crypto lendingSpark is opening access to its $9 billion stablecoin liquidity pool for hedge funds and other institutions to bridge onchain capital with off-chain credit markets.
Read more »

Cardi B and Stefon Diggs spark split rumors after unfollowing each other on social mediaCardi B and Stefon Diggs spark split rumors after unfollowing each other on social mediaThe 'Be Careful' rapper and the NFL star may have called it quits.
Read more »

Cardi B and Stefon Diggs spark split rumors after unfollowing each other on social mediaCardi B and Stefon Diggs spark split rumors after unfollowing each other on social mediaThe 'Be Careful' rapper and the NFL star may have called it quits.
Read more »

Super Bowl spots spark fight over whether we're ready for ads from our chatbotsSuper Bowl spots spark fight over whether we're ready for ads from our chatbotsSuper Bowl ad exposes battle over whether AI chatbots should be delivering ads
Read more »



Render Time: 2026-04-01 17:17:22