For a given level of intelligence, making Mixture of Experts (MOE) models sparser can save a significant amount of flops

Overview Debates Related claims Topics

During the September 2024 inspection of Megaspeed's Malaysian data center, United States officials observed boxed, unopened Nvidia hardware packaged in original plastic

Megaspeed rents out computing power built partially on Nvidia semiconductors to enterprise clients as a remote data center provider

Megaspeed purchased billions of dollars worth of computing equipment from Nvidia

The Qwen 3.6-35B A3B language model contains 40 layers, with each layer comprising 256 experts containing 512 neurons each

On the Artificial Analysis benchmark suite, Fable 5.2 cost $26.54 per task when operating on max effort settings

On the Artificial Analysis Intelligence Index, Fable 5.2 achieved a score of 65, placing 5 points ahead of Gemini Carbon, 6 points ahead of Opus 5.5, and 8 points ahead of GPT-6 Astra

On the RoboDojo Simon robotics benchmark, Fable 5.2 scored 64.9%, exceeding Opus 5.5's score by 25 points

On the Remote Labor Index contractor benchmark, Fable 5.2 achieved a score of 41.3%, compared to Opus 5.5's score of 21%

On protein design benchmarks, Mythos 5.2 scored 85.1%, outperforming Mythos 5.1 by 6 percentage points

On the Artificial Analysis benchmark suite, Fable 5.2 cost $9.08 per task on high effort settings

Ablating 80% to 90% of neurons in the Qwen 3.6-35B A3B language model improved performance on ARC-AGI benchmark puzzles

On the TerminalBench Science 0.1 benchmark for agentic scientific research, Fable 5.2 scored 6 points ahead of GPT-6 Astra

On the Artificial Analysis benchmark suite, Fable 5.2 cost $14.83 per task on extra high effort settings

On the Artificial Analysis Intelligence Index, Fable 5.2 scored 62 on both the Extra High and High effort settings

On the AA Briefcase benchmark, Fable 5.2 recorded a lower Elo score than Opus 5.5

Anthropic priced Fable 5.2 input tokens at $8 per million and output tokens at $40 per million, representing a 20% price reduction compared to Fable 5.1

On the AA Briefcase agentic knowledge work benchmark, Fable 5.2 scored approximately 5 points higher than Fable 5.1, Opus 5.5, and Gemini 4 Carbon

On the FrontierCode coding benchmark, Fable 5.2 scored 4 points ahead of GPT-6 Astra and 3.5 points ahead of Opus 5.5

Nvidia designs and sells semiconductor chips utilized for artificial intelligence workloads in data centers

Selecting the top 64 activating neurons per expert in Qwen 3.6-35B A3B ablates approximately 88% of all neuron activations