SSyncDev

Llama 3.1 405B

Meta AI·Text → Text·active
Compare models

Meta's largest open-weights model, competitive with frontier closed models.

Key specifications

Intelligence
74.0
Context window
128K
Max output
4.1K
Output speed
30 t/s
Latency (TTFT)
0.70s
Input $/1M
$3.50
Output $/1M
$3.50
License
Llama 3.1 Community
Architecture
Transformer (dense)
Parameters
405B

Capabilities

Function callingStreamingJSON modeFine-tuningOpen weights

Strengths

  • + Open weights — self-hostable
  • + Strong general performance

Limitations

  • − Heavy to run
  • − Slower hosted throughput

Benchmark scores

BenchmarkScore
GPQA Diamond51.1%
MMLU-Pro73.3%
SWE-Bench Verified24.0%
HumanEval89.0
AIME 20249.0%