Anthropic · OpenAI · Google
Flagship
No data
Cheapest
No pricing data
Fastest
No speed data
Best Value
No value data
Recent Moves
No recent changes
Flagship
GPT-5.2
AA Index: 47 · Rank #2
Cheapest
GPT-5.1
$1.13/M (1:1 blend)
Fastest
GPT-5.1
118 tok/s · 5.6s TTFT
Best Value
GPT-5.1
Value score: 1294
Recent Moves
No recent changes
Flagship
No data
Cheapest
No pricing data
Fastest
No speed data
Best Value
No value data
Recent Moves
No recent changes
Daily domain rank from Tranco (aggregated DNS/web traffic). Lower rank = more popular. 30-day rank change is the key signal — improvement indicates growing adoption or a viral moment.
Tracks median inference API cost across providers. Frontier = proprietary models (OpenAI, Anthropic, Google, DeepSeek, Mistral, xAI, Cohere). Open = open-weight models (Llama, Qwen, Mixtral, etc.).
Daily PyPI download velocity for GPU-specific packages. GPU Deploy packages proxy real GPU utilization. Training packages track distributed training workload intensity. Inference Demand packages are a leading indicator for GPU capacity needs.
Daily Docker Hub pull velocity for GPU container images. GPU Runtime pulls proxy real deployment activity. Training image pulls track distributed training workload spin-ups. Inference pulls measure production serving deployment pace.
Data sources: Artificial Analysis (capability, pricing, speed), LMArena (human preference arena scores).