📊 Llama 3.1 Tulu3 405B’s independent benchmark run: GPQA 51.6%, MMLU-Pro 71.6%, HLE 3.3%, LiveCodeBench 29.1%. The gap between reasoning and coding tells the real story — see how it stacks up on the full leaderboard.

https://olud.ai/leaderboard.html

#LLM #Benchmarks #OpenSource #AI

Cua looks like a real bet on the boring parts of agent infrastructure
https://papoo.work/doc/a6baceaaed1f9be9
#claudenews #agents #computer_use #benchmarks

Cua looks like a real bet on the boring parts of agent infrastructure
https://papoo.work/doc/a6baceaaed1f9be9
#claudenews #agents #computer_use #benchmarks

Cua looks like a real bet on the boring parts of agent infrastructure
https://papoo.work/doc/a6baceaaed1f9be9
#claudenews #agents #computer_use #benchmarks

Cua looks like a real bet on the boring parts of agent infrastructure
https://papoo.work/doc/a6baceaaed1f9be9
#claudenews #agents #computer_use #benchmarks

Cua looks like a real bet on the boring parts of agent infrastructure
https://papoo.work/doc/a6baceaaed1f9be9
#claudenews #agents #computer_use #benchmarks

We benchmarked the iPhone 18 Pro Max against the Galaxy S26 Ultra. The A20 Pro leads by ~30% in single-core and ~9% in multi-core, and pulls far ahead in sustained gaming tests. Full results: ag4tech.com/iphone-18-pr...

#iPhone18ProMax #A20Pro #Benchmarks

Cua looks like a real bet on the boring parts of agent infrastructure
https://papoo.work/doc/a6baceaaed1f9be9
#claudenews #agents #computer_use #benchmarks

📊 Mistral Small 4 (Non-reasoning) — the actual numbers

GPQA: 57.1%
Humanity's Last Exam: 3.8%
Long Context Reasoning: 28.3%
IFBench: 32.8%

⚡ 146.5 tokens/sec
💰 34.4 intelligence points per dollar

Measured independently,
https://olud.ai/leaderboard.html

#LLM #Benchmarks #OpenSource #AI

📊 Nova 2.0 Omni (Non-reasoning) — the actual numbers

GPQA: 55.5%
MMLU-Pro: 71.9%
Humanity's Last Exam: 3.8%
Long Context Reasoning: 24.7%

💰 9.6 intelligence points per dollar

Measured independently, not self-reported →https://olud.ai/leaderboard.html

#LLM #Benchmarks #OpenSource #AI

SPOTTED: Two unnamed models, dxzui and clzui, appeared on LMArena leaderboard without vendor attribution, suggesting unreleased codenames under testing.
Single source: LMArena. Unconfirmed.

#AI #benchmarks

ON THE RADAR: Anthropic appoints Accenture as first embedded evaluator to implement safety slowdown proposal.

#AI #benchmarks

M5 Ultra slashes past M3 Ultra by nearly 60% in Metal—M6 beats M5 by ~34%. #Apple #GPU #M5Ultra #M6 #Metal #GraphicsPerformance #Benchmarks https://thedailytechfeed.com/m5-ultra-m6-gpu-benchmarks-drop-major-gains-vs-m3/

FRONTIER LEADERBOARD: the strongest proprietary AI models right now.
1. claude-opus-4-6 — anthropic
2. claude-opus-4-5 — anthropic
3. gemini-3.6-flash — google
4. claude-sonnet-4-5 — anthropic
5. gemini-3.1-pro
Ranked by LMArena human votes.

#AI #benchmarks

📊 Nova 2.0 Lite (Non-reasoning) — the actual numbers

GPQA: 60.3%
MMLU-Pro: 74.3%
Humanity's Last Exam: 2.9%
Long Context Reasoning: 18.7%

⚡ 226.5 tokens/sec
💰 10.2 intelligence points per dollar

Measured independently, n
https://olud.ai/leaderboard.html

#LLM #Benchmarks #OpenSource #AI

SPOTTED: Unnamed model "lufei" appeared on LMArena leaderboard with no vendor attribution, likely an unreleased codename.
Single source: LMArena. Unconfirmed.

#AI #benchmarks

The AI landscape just got a major earthquake! Claude Opus 5 isn't just a new contender; it's delivering impressive benchmarks and pricing that

#Claude #ChatGPT #Benchmarks #Pricing #Token

MILESTONE: Shanghai BO Laboratory's 744B-parameter Atria Dawn ranked first on five of sixteen benchmarks, topping GPT-5.6 Sol on web search and potato tasks.

#AI #benchmarks

MILESTONE: Shanghai AI Laboratory's 744B-parameter Atria Dawn ranked first on five of sixteen benchmarks, topping GPT-5.6 Sol on web search and cyber tasks.

#AI #benchmarks

MLPerf Inference v6.1: Intel Xeon & Arc lebt, 512 MI355X skalieren und VR200 debütiert #MLPerf #Benchmarks