---
id: provider-latency/2026-07
series: provider-latency
title: Time to first token by provider, July 2026
period: 2026-07
period_start: 2026-07-01
period_end: 2026-07-31
updated: 2026-11-11
license: CC BY 4.0
canonical: https://www.requesty.ai/data/provider-latency/2026-07
latest: https://www.requesty.ai/data/provider-latency
---

# Time to first token by provider, July 2026

> Sference had the fastest median time to first token among major providers in July 2026 at 46 ms. The median major provider took 1.20 s.

Question: Which LLM provider returns the first token fastest?

## Key findings

1. Fastest median TTFT: Sference at 46 ms, then Nebius at 463 ms.
2. Highest median output speed: Nebius at 196 tokens per second.
3. Fastest median full streamed response: Azure at 1.27 s.
4. OpenAI moved the most since June 2026: median TTFT went from 3.48 s to 923 ms (-2.56 s).

## Data

| Provider | Median TTFT | p90 TTFT | Median total | Median tokens/s | Streamed share | Organizations (band) |
| --- | --- | --- | --- | --- | --- | --- |
| Sference | 46 ms | 87 ms | 1.62 s | 89.800 | 89.71% | 100-499 |
| Nebius | 463 ms | 4.04 s | 1.65 s | 196.200 | 46.92% | 500+ |
| Azure | 689 ms | 2.08 s | 1.27 s | 85.900 | 63.68% | 500+ |
| DeepInfra | 699 ms | 2.45 s | 8.25 s | 22.800 | 58.59% | 500+ |
| DeepSeek | 834 ms | 1.65 s | 5.04 s | 49.800 | 18.17% | 500+ |
| OpenAI | 923 ms | 8.95 s | 2.24 s | 120.100 | 6.17% | 500+ |
| xAI | 970 ms | 2.36 s | 9.63 s | 33.600 | 26.84% | 100-499 |
| Vertex (Claude) | 1.19 s | 3.02 s | 5.39 s | 91.800 | 68.93% | 500+ |
| Fireworks | 1.22 s | 4.83 s | 5.43 s | 108.400 | 51.65% | 500+ |
| Mistral | 1.25 s | 13.5 s | 2.99 s | 115.400 | 27.02% | 500+ |
| TensorX | 1.38 s | 7.53 s | 4.42 s | 94.500 | 54.32% | 500+ |
| Bedrock | 1.72 s | 4.13 s | 5.70 s | 95.100 | 34.58% | 500+ |
| Alibaba | 1.78 s | 3.94 s | 5.01 s | 44.600 | 14.46% | 100-499 |
| Vertex (Gemini) | 1.80 s | 8.39 s | 3.50 s | 111.800 | 8.16% | 500+ |
| Anthropic | 1.96 s | 3.77 s | 6.85 s | 80.100 | 51.69% | 500+ |
| Novita | 3.71 s | 11.4 s | 9.99 s | 53.000 | 7.43% | 500+ |

## Method

- Time to first token and tokens per second are measured on streamed, successful requests only, with the same formulas /rankings uses. Tokens per second counts visible output tokens after the first token, with a 250 ms floor on decode time and at least 20 output tokens.
- A provider needs at least 500 streamed successes in the month for latency to be published. Only providers with at least 0.5% of published traffic are shown.
- A row is published only when it has at least 20 requests, at least 5 distinct organizations, and no single organization above 80% of its traffic. These are the same floors /rankings uses. Rows below the floors are withheld, not estimated.
- Only shares, rates, medians and organization bands are published. Absolute request, token and spend totals stay private.

## Caveats

- Latency depends on the model mix. A provider serving mostly small models will look faster than one serving large reasoning models.
- Reasoning models think before the first visible token, which raises TTFT.

## Cite as

Requesty (2026). Time to first token by provider, July 2026. Requesty Data. https://www.requesty.ai/data/provider-latency/2026-07

```bibtex
@misc{requesty_provider_latency_2026_07,
  author       = {{Requesty}},
  title        = {Time to first token by provider, July 2026},
  year         = {2026},
  howpublished = {\url{https://www.requesty.ai/data/provider-latency/2026-07}},
  note         = {Requesty Data}
}
```

Source: Requesty production gateway. License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/).
Downloads: [JSON](https://www.requesty.ai/data/provider-latency/2026-07/data.json) · [CSV](https://www.requesty.ai/data/provider-latency/2026-07/data.csv) · [Markdown](https://www.requesty.ai/data/provider-latency/2026-07/data.md)