together Provider Benchmarks

Comprehensive performance summary covering 11 models.

This provider hub highlights throughput and latency trends across every together model monitored by LLM Benchmarks. Use it to compare hosting tiers, track regressions, and discover the fastest variants in the catalogue.

Visit together Official Website

Provider Snapshot

Models Tracked

11

Median Tokens / Second

73.51

Median Time to First Token (ms)

0.61

Last Updated

Dec 25, 2025

Key Takeaways

  • 11 together models are actively benchmarked with 11895 total measurements across 10698 benchmark runs.

  • llama-3.1-8b leads the fleet with 150.00 tokens/second, while llama-3.3-70b delivers 66.70 tok/s.

  • Performance varies by 124.9% across the together model lineup, indicating diverse optimization strategies for different use cases.

  • Median time to first token across the fleet is 0.61 ms, showing excellent responsiveness for interactive applications.

  • The together model fleet shows varied performance characteristics (53.4% variation coefficient), reflecting diverse model architectures.

Fastest Models

ProviderModelAvg Toks/SecMinMaxAvg TTF (ms)
togetherllama-3.1-8b150.002.14261.000.28
togethermistral-7b124.008.16172.000.28
togetherllama-3.2-3b104.005.04148.000.37
togetherqwen-2.5-7b98.607.97145.000.21
togetherllama-3.1-70b73.702.11152.000.44
togetherllama-3.3-70b66.701.11128.000.72

All Models

Complete list of all together models tracked in the benchmark system. Click any model name to view detailed performance data.

ProviderModelAvg Toks/SecMinMaxAvg TTF (ms)
togetherllama-3.1-8b150.002.14261.000.28
togetherllama-3.1-70b73.702.11152.000.44
togetherllama-3.1-405b22.003.3330.200.65
togetherllama-3.2-3b104.005.04148.000.37
togethermistral-7b124.008.16172.000.28
togethermixtral-8x7b49.202.26109.000.88
togetherqwen-2.5-7b98.607.97145.000.21
togetherqwen-2.5-72b55.201.1669.600.47
togetherdeepseek-v327.701.8065.901.01
togetherdeepseek-r137.501.5067.201.35
togetherllama-3.3-70b66.701.11128.000.72

Featured Models

Frequently Asked Questions

Based on recent tests, llama-3.1-8b shows the highest average throughput among tracked together models.

This provider summary aggregates 11895 individual prompts measured across 10698 monitoring runs over the past month.