HomeTagsBenchmarks

Benchmarks - Page 3

Latest

Shanghai AI Lab Releases Atria Dawn Open Weights With 512K Context

Shanghai AI Lab has posted full Atria Dawn weights on Hugging Face under MIT, with 512K context. Its own table has it trailing Opus 5 on SWE-bench Pro.

Microsoft Diagnostic AI Scored 83% on Past Cases, Science Paper Says

Microsoft's MAI-DxO diagnosed 83% of complex past cases against 36% for doctors in a Science benchmark study. No product, approval or date is announced.

Mistral Launches Voxtral 3 Speech Models and a Voice Agent SDK

Mistral's Voxtral 3 adds speech-to-text in 50 languages, expressive TTS and a LiveKit-based voice agent SDK. Benchmark wins are Mistral's own claims.

ElevenLabs Launches Eleven Realtime Speech-to-Speech Models for Voice Agents

ElevenLabs launches Eleven Realtime v1 Mini and Max, speech-to-speech voice agent models with 90+ languages and an OpenAI-compatible API.

OpenAI Says Its Latest Models Are Far Better at COBOL and Fortran Code

OpenAI says GPT-6 Astra and GPT-6.1 Sol score far higher on its own COBOL and Fortran coding evals, though production use still needs expert review.

Epoch AI Says Open-Weight Models Trail Closed Ones on Cyber Skills by 6-7 Months

Epoch AI estimates open-weight models trail closed ones on cyber skills by 6-7 months, with Z.ai's GLM-5.3 on top, but its score rests on three results.

DuckDB Posts Unaudited TPC-H SF1,000 Results for v2.0-dev on a Single NVIDIA Vera Socket

DuckDB published unaudited TPC-H SF1,000 results for a v2.0-dev preview on one NVIDIA Vera socket. Figures are DuckDB’s own, with NVIDIA’s help.

Google Releases Gemini 4 Argon Worldwide: What It Costs, Who Gets It and What Google Claims

Google has made Gemini 4 Argon available worldwide. API prices are $2/$10 per million tokens until 30 November, then $4/$20. Free users don't get it.

Seqera Launches Nextflow Frontier, an AI System That Runs Thousands of Biology Experiments

Seqera has launched Nextflow Frontier, an AI system that runs thousands of biology experiments. Its benchmark results are Seqera's own claims.

Anthropic Says New Defences Cut the Data “Distillers” Harvest From Claude by Around 90 to 99 Percent

Anthropic says new defences cut the data harvested by distillers copying Claude by around 90 to 99 percent, citing early evidence. Here is how they work.

Latest on TSN

Explore the TSN Atlas: Every topic we cover, mapped. Click a bubble to dive in.