GPT-5
GPT-5 is a reasoning model from OpenAI, released in the API on August 7, 2025 (snapshot gpt-5-2025-08-07). It takes text and image input, returns text, and has configurable reasoning effort.
Deprecated. OpenAI will shut down gpt-5-2025-08-07 in the API on December 11, 2026 (announced June 11, 2026) and names gpt-5.6-sol as its replacement. Its model page calls GPT-5 a previous model and recommends GPT-6 Astra.
Released
August 7, 2025
Type
reasoning
License
proprietary
Context
400,000 tokens
Up to 272,000 input tokens and 128,000 output tokens.
Pricing
Input
$1.25 / 1M tokens
Output
$10.00 / 1M tokens
List price on the developer's own pages, last checked , more than 7 days ago: it may have changed since.
Benchmarks
| SWE-bench VerifiedOpenAI’s figure, on 477 of the 500 tasks. | 74.9% |
| Aider polyglotOpenAI’s figure. | 88% |
Scores are the developer's own published figures for this model, under the test conditions noted. Benchmarks with no primary source are not listed.
Capabilities
Sources
Every figure on this page was checked against these pages on . A figure we could not source is left out.
- OpenAI API docs: GPT-5 (opens in a new tab)
- OpenAI API docs: Deprecations (opens in a new tab)
- OpenAI: Introducing GPT-5 for developers (opens in a new tab)
Spotted something out of date? Tell us.
Links
In the news: GPT-5 and OpenAI
OpenAI
An updated Model Spec for more capable and beneficial AI
OpenAI has released an updated Model Spec that outlines five core values. These values are intended to guide the development of more capable AI systems, ensuring they exhibit sound judgment, uphold human agency, and remain faithful to human intent.
The Hacker News
Fake ChatGPT and Claude Apps Abuse Signed Executables to Deploy Crypto Miners
Threat actors are distributing malicious loaders and cryptocurrency miners disguised as artificial intelligence (AI) desktop applications for Anthropic Claude and OpenAI ChatGPT. This campaign has been dubbed Yoba. The activity involves the creation of fake GitHub projects that falsely claim to be desktop applications such as GPT-5.6 Sol, GPT-6 Astra, and Claude Opus. These fake applications are used to deliver a malware family. These malicious applications abuse signed executables to deploy cryptocurrency miners, posing a security risk to users seeking AI tools.
MarkTechPost
Infini-AI-Lab Releases Hermit, an Open-Source Harness Adapter Framework for AI Agents
Infini-AI-Lab has released Hermit, an open-source harness adapter framework designed for AI agents. This framework allows for the training of an external 'harness adapter' while the agent's core model and harness remain unchanged. The first version, v0.1.0, was released on October 25, 2026. CMU postdoc Haizhong Zheng announced the release on X on October 27, 2026. Hermit aims to enable developers to adapt agents like Codex or Claude Code to new tasks without requiring fine-tuning or forking the harness. Hermit is a Python framework with no parameters of its own and is available under an Apache-2.0 license, installable via pip. The offline Quick Start version requires no GPU, model account, or dataset. For actual use, it can run with any API backend or a local vLLM, SGLang, or Ollama server. Performance reports indicate gains over baselines without adapters on benchmarks like OfficeQA, CL-Bench, and ServeLearnBench, using GLM 5.3 Flash. The MWM adapter achieved 0.795 accuracy on CL-Bench database exploration, surpassing the in-context baseline of 0.580. However, the prompt-only adapter scored 0.470 on the same benchmark. A harness adapter functions as a trainable external component, similar to a plugin with integrated training support. Unlike model adapters like LoRA that adapt a model, a harness adapter adapts an existing agent. It can incorporate instructions, skills, memory, tools, executable code, and agent coordination logic. The design choice to keep the base model and harness static is attributed to several factors: many harnesses are closed-source, editing open-source ones can disrupt fine-tuned component interactions, and a learned adapter can be transferred between compatible models and harnesses. Hermit employs an RL-style training loop for adapter training.
MarkTechPost
Block Unveils Dessa, a Full-Duplex Speech Model Trained Only on Speech
Block's AI research lab has released Dessa, a full-duplex speech model trained exclusively on human conversation, announced as a research preview on October 23, 2026. Dessa functions as a speech-to-speech model, processing audio input and producing audio output, enabling it to listen and talk concurrently. The model was trained without relying on text data or pre-existing models, diverging from the text-first approach common in current voice agents. Block states that Dessa learned conversational nuances, such as when to interject, use backchannels like "yeah," and incorporate natural interruptions like laughter. The training process for Dessa involved only human conversation audio, with no text or pre-trained models utilized. Block claims this method allows Dessa to learn language with significantly fewer words than required by LLMs. Inference for Dessa runs faster than real-time on hardware like a MacBook Pro M4 Max or superior. The model was trained on one NVIDIA GB200 NVL72 rack, which comprises 72 Blackwell GPUs and 36 Grace CPUs. Block's approach aims to preserve emotional cues, sarcasm, laughter, and sighs that may be lost in traditional speech recognition, text modeling, and speech synthesis chains.
Search Engine Journal
ChatGPT Data: 1 In 5 Consumer Messages Now Seek Information via @sejournal, @MattGSouthern
Consumer messages to ChatGPT seeking information now constitute 20% of all consumer messages. This usage level matches the share of messages dedicated to writing tasks. The increase in information-seeking messages is smaller than observed in earlier data.