---
title: "Fastest LLM API for Classification Routing 2026 | Openbenchmarks"
description: "Fastest LLM API for classification and ticket routing: end-to-end p99 on 200 ticket-triage calls per provider across 10 providers. Open source."
url: https://openbenchmarks.com/inference/fastest-inference-provider-for-classification
---

[](https://openbenchmarks.com/) 

[← LLM inference provider benchmark](https://openbenchmarks.com/inference) 

[benchmarks](https://openbenchmarks.com/) / [inference](https://openbenchmarks.com/inference) / fastest inference provider for classification 

classification latency · ticket routing · measured tail · measured Sep 3, 2026

# Fastest LLM API for classification and ticket routing

**What the benchmark found.** Baseten recorded the lowest end-to-end p99 latency at 1.20 seconds on this run, with a median of 274 ms, time to first token p95 of 538 ms, task success of 98.0%, and failure rate of 0.0%. The ordering is computed from the ticket-triage family alone, with 200 submitted classifications per provider. Measured across 10 LLM inference providers serving GLM 5.3 Flash, 200 requests each, Sep 3, 2026.

**Providers measured.** Baseten, DeepInfra, Fireworks AI, Modal, Nebius, Novita AI, Parasail, Telnyx, Together AI, and Z.AI, each serving GLM 5.3 Flash behind an OpenAI-compatible chat-completions endpoint.

**What this page compares.** A synchronous classifier blocks the next application branch until the complete category and priority object arrives. This page uses only the 200 ticket-triage requests per provider and orders them by complete-response end-to-end p99 rather than pooled benchmark latency.

**How to read the result.** Use end-to-end p95 for the latency most requests must fit and p99 for the rarer waits that can breach a routing SLA. Read both beside exact task success and failure rate: a fast wrong label or failed request does not complete the routing job.

**Workload and configuration.** Every provider received the same 600 deliberately easy structured tasks: 200 meeting-note lookups, 200 ticket-triage classifications, and 200 contract-term extractions. GLM 5.3 Flash ran with streaming enabled, reasoning low, temperature 0, top-p 1, a 256-token output ceiling, one attempt, a 20-second timeout, and per-provider concurrency one.

**Task-specific table.** The table below uses only the published `ticket-triage` aggregate rather than the three-family pooled summary. Each provider therefore contributes 200 submitted requests to this view.

**APIs used.** The runner called each provider's OpenAI-compatible chat-completions surface with the model identifiers below; the links open the provider documentation used to verify each adapter.

* **Baseten:** `zai-org/GLM-5.3-Flash` · [API docs ↗](https://docs.baseten.co/reference/inference-api/chat-completions)
* **Modal:** `zai-org/GLM-5.3-Flash` · [API docs ↗](https://modal.com/docs/guide/endpoints#calling-your-endpoint)
* **Telnyx:** `zai-org/GLM-5.3-Flash` · [API docs ↗](https://developers.telnyx.com/api-reference/openai-chat/create-a-chat-completion-openai-compatible)
* **Novita AI:** `zai-org/glm-5.3-flash` · [API docs ↗](https://novita.ai/docs/api-reference/model-apis-llm-create-chat-completion)
* **Z.AI:** `glm-5.3-flash` · [API docs ↗](https://docs.z.ai/api-reference/llm/chat-completion)
* **Together AI:** `zai-org/GLM-5.3-Flash` · [API docs ↗](https://docs.together.ai/docs/inference/chat/overview)
* **Fireworks AI:** `accounts/fireworks/models/glm-5p3-flash` · [API docs ↗](https://docs.fireworks.ai/api-reference/post-chatcompletions)
* **Parasail:** `zai-org/GLM-5.3-Flash` · [API docs ↗](https://docs.parasail.io/parasail-docs/api-reference/chat-completions)
* **Nebius:** `zai-org/GLM-5.3-Flash` · [API docs ↗](https://docs.tokenfactory.nebius.com/api-reference/inference/create-chat-completion)
* **DeepInfra:** `zai-org/GLM-5.3-Flash` · [API docs ↗](https://docs.deepinfra.com/api-reference/chat-completions/openai-chat-completions)

\[01\] measured comparison

## Which LLM API returns a classification fastest at p99?

Lower latency is better. The delta is calculated from the fastest ticket-triage p99, while p50 and p95 show whether the same provider also leads away from the extreme tail.

benchmarks/inference/fastest-inference-provider-for-classification reviewed run 

__LLM inference providers sorted by end-to-end p99 latency__
| Provider | End-to-end p50 | End-to-end p95 | End-to-end p99 | Δ vs fastest p99 | Task success | Failure rate |
| --- | --- | --- | --- | --- | --- | --- |
| [Baseten](https://www.baseten.co "Open Baseten in a new tab") [zai-org/GLM-5.3-Flash ↗](https://docs.baseten.co/reference/inference-api/chat-completions) | 274 ms | 749 ms | 1.20 s | 0 ms | 98.0% | 0.0% |
| [Modal](https://modal.com/ "Open Modal in a new tab") [zai-org/GLM-5.3-Flash ↗](https://modal.com/docs/guide/endpoints#calling-your-endpoint) | 493 ms | 1.21 s | 2.76 s | +1555 ms | 98.0% | 0.0% |
| [Telnyx](https://telnyx.com "Open Telnyx in a new tab") [zai-org/GLM-5.3-Flash ↗](https://developers.telnyx.com/api-reference/openai-chat/create-a-chat-completion-openai-compatible) | 645 ms | 2.24 s | 3.78 s | +2582 ms | 98.5% | 0.0% |
| Novita AI [zai-org/glm-5.3-flash ↗](https://novita.ai/docs/api-reference/model-apis-llm-create-chat-completion) | 1.63 s | 2.77 s | 4.11 s | +2905 ms | 95.0% | 3.5% |
| Z.AI [glm-5.3-flash ↗](https://docs.z.ai/api-reference/llm/chat-completion) | 1.62 s | 3.32 s | 5.75 s | +4551 ms | 96.5% | 0.5% |
| [Together AI](https://www.together.ai "Open Together AI in a new tab") [zai-org/GLM-5.3-Flash ↗](https://docs.together.ai/docs/inference/chat/overview) | 417 ms | 1.48 s | 6.83 s | +5629 ms | 98.5% | 0.0% |
| Fireworks AI [accounts/fireworks/models/glm-5p3-flash ↗](https://docs.fireworks.ai/api-reference/post-chatcompletions) | 1.07 s | 5.71 s | 7.70 s | +6498 ms | 98.5% | 0.0% |
| Parasail [zai-org/GLM-5.3-Flash ↗](https://docs.parasail.io/parasail-docs/api-reference/chat-completions) | 1.48 s | 6.08 s | 13.1 s | +11924 ms | 98.5% | 0.5% |
| Nebius [zai-org/GLM-5.3-Flash ↗](https://docs.tokenfactory.nebius.com/api-reference/inference/create-chat-completion) | 1.02 s | 6.19 s | 15.0 s | +13789 ms | 84.5% | 14.5% |
| DeepInfra [zai-org/GLM-5.3-Flash ↗](https://docs.deepinfra.com/api-reference/chat-completions/openai-chat-completions) | 1.23 s | 8.43 s | 17.1 s | +15870 ms | 96.5% | 2.0% |

### Provider by provider, in the order above

* Baseten: 274 ms median, 1.20 seconds end-to-end p99, time to first token p95 538 ms, 313.0 tokens per second, 98.0% task success, 0.0% failures.
* Modal: 493 ms median, 2.76 seconds end-to-end p99, time to first token p95 1.06 seconds, 270.3 tokens per second, 98.0% task success, 0.0% failures.
* Telnyx: 645 ms median, 3.78 seconds end-to-end p99, time to first token p95 2.21 seconds, 211.9 tokens per second, 98.5% task success, 0.0% failures.
* Novita AI: 1.63 seconds median, 4.11 seconds end-to-end p99, time to first token p95 2.64 seconds, 65.4 tokens per second, 95.0% task success, 3.5% failures.
* Z.AI: 1.62 seconds median, 5.75 seconds end-to-end p99, time to first token p95 2.65 seconds, 57.2 tokens per second, 96.5% task success, 0.5% failures.
* Together AI: 417 ms median, 6.83 seconds end-to-end p99, time to first token p95 1.31 seconds, 178.6 tokens per second, 98.5% task success, 0.0% failures.
* Fireworks AI: 1.07 seconds median, 7.70 seconds end-to-end p99, time to first token p95 5.62 seconds, 63.3 tokens per second, 98.5% task success, 0.0% failures.
* Parasail: 1.48 seconds median, 13.12 seconds end-to-end p99, time to first token p95 5.67 seconds, 48.8 tokens per second, 98.5% task success, 0.5% failures.
* Nebius: 1.02 seconds median, 14.99 seconds end-to-end p99, time to first token p95 6.10 seconds, 323.3 tokens per second, 84.5% task success, 14.5% failures.
* DeepInfra: 1.23 seconds median, 17.07 seconds end-to-end p99, time to first token p95 6.23 seconds, 89.5 tokens per second, 96.5% task success, 2.0% failures.

\[02\] workflow fit

## Where is classification latency blocking?

* **Pre-chat intent detection.** The classifier must choose the bot flow or human queue before the interaction can continue.
* **Incident severity routing.** A priority enum controls paging or escalation, so completion latency and exactness must be evaluated together.
* **Tool dispatch.** A bounded class selects the next API call; end-to-end latency, not first-token latency, determines when dispatch can occur.

\[03\] how it was measured

## How is LLM inference provider latency and accuracy measured here?

* **Same model.** Every endpoint serves GLM 5.3 Flash; the provider layer is the only variable.
* **Same request.** One-turn, non-tool chat completion in JSON-object mode, streaming on, temperature 0, top-p 1, reasoning low, 256-token ceiling, 20-second timeout, concurrency one.
* **Client clock, pinned tokenizer.** Time to first token, end-to-end latency and tokens per second come from the client; visible answer tokens are counted with a pinned `tiktoken o200k_base` tokenizer, never provider-reported usage.
* **Exact task success.** An answer counts only when it is schema-valid and exactly correct under the task rubric. Timeouts, HTTP errors and malformed responses stay in the denominator.
* **Serverless endpoints only.** Dedicated capacity is outside the comparison. Runner and per-request records: [openbenchmarks-labs/inference ↗](https://github.com/openbenchmarks-labs/inference).

\[04\] faq

## Fast classification APIs: ticket-family evidence

### Which provider had the lowest ticket-classification end-to-end p99?

Baseten recorded the lowest end-to-end p99 latency at 1.20 seconds on this run, with a median of 274 ms, time to first token p95 of 538 ms, task success of 98.0%, and failure rate of 0.0%. The ordering is computed from the ticket-triage family alone, with 200 submitted classifications per provider.

### Did one provider lead both median and tail ticket-routing latency?

Baseten had the lowest median end-to-end latency at 274 ms, while Baseten had the lowest end-to-end p99 at 1.20 seconds. The same provider led both points of the distribution.

### How do timed-out classification calls affect this comparison?

Timeouts, HTTP errors, and transport failures have no completed visible answer and therefore no honest end-to-end latency value. They stay in the submitted-request denominator for failure rate and task success, which must be read beside the latency columns.

\[05\] related comparisons

## Same run, other questions

* [LLM inference provider benchmark: complete scorecard and methodology →](https://openbenchmarks.com/inference)
* [Best LLM API for text classification and ticket routing →](https://openbenchmarks.com/inference/best-inference-provider-for-classification)
* [Best LLM API for intent routing and ticket triage →](https://openbenchmarks.com/inference/best-inference-provider-for-intent-routing)
* [Best LLM API for customer support automation →](https://openbenchmarks.com/inference/best-inference-provider-for-customer-support-automation)
* [Best LLM inference provider for small tasks and short prompts →](https://openbenchmarks.com/inference/best-inference-provider-for-small-tasks)
* [Fastest LLM inference provider for small tasks →](https://openbenchmarks.com/inference/fastest-inference-provider-for-small-tasks)
* [Best LLM inference provider for enterprise AI programs →](https://openbenchmarks.com/inference/best-inference-provider-for-ai-transformation)

\[06\] source benchmark

## Read the complete LLM inference provider benchmark

Metric definitions, request scheduling, task rubrics, failure handling, and the complete provider scorecard live on the [LLM Inference Provider Benchmark →](https://openbenchmarks.com/inference) 

Openbenchmarks [inference benchmark](https://openbenchmarks.com/inference) [all benchmarks](https://openbenchmarks.com/) © 2026 Openbenchmarks