evald.ai

METR Blog · April 10, 2026 07:00

MirrorCode: Evidence that AI can already do some weeks-long coding tasks

This is a linkpost for MirrorCode, a project that METR funded and co-developed with Epoch AI. See Epoch AI’s blog post for more detail: https://epoch.ai/blog/mirrorcode-preliminary-results/

OpenAI Evaluation Filter · April 10, 2026 00:00

Creating images with ChatGPT

Learn how to create and refine images with ChatGPT using clear prompts, iterate on designs, and generate high-quality visuals in minutes.

Topics: LLM Evaluation

Entities: LLM EvaluationChatGPT

OpenAI Evaluation Filter · April 10, 2026 00:00

Using skills

Learn how to create and use ChatGPT skills to build reusable workflows, automate recurring tasks, and ensure consistent, high-quality outputs.

Topics: LLM Evaluation

Entities: LLM EvaluationChatGPT

OpenAI Evaluation Filter · April 09, 2026 00:00

CyberAgent moves faster with ChatGPT Enterprise and Codex

CyberAgent uses ChatGPT Enterprise and Codex to securely scale AI adoption, improve quality, and accelerate decisions across advertising, media, and gaming.

Topics: LLM Evaluation

Entities: LLM EvaluationChatGPT

MLCommons Evaluation Filter · April 06, 2026 14:57

MLCommons Releases MLPerf Client v1.6 with Performance Optimizations and Enhanced User Experience - MLCommons

MLCommons releases MLPerf Client v1.6 with updated Windows ML and llama.cpp support, Apple MLX improvements for Mac and iPad, and usability enhancements for faster, more reliable AI benchmarking on personal computers.

Topics: Benchmarks

Entities: Benchmarks

MLCommons Evaluation Filter · April 01, 2026 14:50

MLCommons Releases New MLPerf Inference v6.0 Benchmark Results - MLCommons

MLCommons releases MLPerf Inference v6.0 results — the most significant benchmark update to date, with new tests for text-to-video, GPT-OSS 120B, DLRMv3, vision-language models, and YOLOv11

Topics: Benchmarks

Entities: Benchmarks

METR Blog · April 01, 2026 07:00

Fine-tuning experiments on CoT controllability

We find that a small amount of fine-tuning on instruction following in the CoT generalizes to meaningful increases in CoT controllability on an out-of-distribution set of tasks. We fine-tune four reasoning models on small datasets of instruction-following...