LLM Gateways · week of Apr 13, 2026

LLM Gateways leaderboard

Scores are shown as X/100 · #N of M. WoW delta compares this week to the previous completed run.

#	Brand	Score	Mention	Pos	Coverage	WoW
1	LiteLLM	50.0/100 #1 of 8	28.0%	0.67	80.0%	—
2	LangSmith	49.0/100 #2 of 8	20.0%	0.90	60.0%	—
3	Langfuse	43.0/100 #3 of 8	20.0%	0.71	60.0%	—
4	Helicone	39.0/100 #4 of 8	20.0%	0.57	60.0%	—
5	Portkey	36.0/100 #5 of 8	12.0%	0.59	60.0%	—
6	OpenRouter	35.0/100 #6 of 8	16.0%	0.63	40.0%	—
7	Vercel AI Gateway	30.0/100 #7 of 8	12.0%	0.67	20.0%	—
8	Braintrust	29.0/100 #8 of 8	8.0%	0.56	40.0%	—

Brand × engine breakdown

Each cell shows mention count in this category and average rank when present.

Brand	PPLX	GPT	CLD	GEM	DSK
LiteLLM	2 · #2.0	0	2 · #1.0	1	2
LangSmith	2 · #1.0	0	1 · #1.0	0	2 · #1.0
Langfuse	2 · #3.5	0	1 · #1.0	2 · #3.5	0
Helicone	0	0	2 · #2.0	1 · #5.0	2
Portkey	1 · #2.0	0	1 · #3.0	0	1
OpenRouter	2 · #1.0	0	0	0	2
Vercel AI Gateway	3 · #2.5	0	0	0	0
Braintrust	1 · #2.0	0	1 · #6.0	0	0

Prompt wins and gaps

Winning prompts are 5/5 engine hits. Losing prompts are 0/5 (or lowest coverage fallback).

LiteLLM

50.0/100 · #1 of 8

Winning prompts (5/5 engines)

No full-coverage prompts this week.

Losing prompts (0/5 engines)

LLM observability platform for tracking prompts, costs, and latency
OpenAI proxy with caching and rate limiting
LLM analytics tool for debugging agent workflows

LangSmith

49.0/100 · #2 of 8

Winning prompts (5/5 engines)

No full-coverage prompts this week.

Losing prompts (0/5 engines)

Best LLM gateway for routing across multiple model providers
OpenAI proxy with caching and rate limiting
Router that fails over between Claude, GPT, and open-source models

Langfuse

43.0/100 · #3 of 8

Winning prompts (5/5 engines)

No full-coverage prompts this week.

Losing prompts (0/5 engines)

Best LLM gateway for routing across multiple model providers
OpenAI proxy with caching and rate limiting
Router that fails over between Claude, GPT, and open-source models

Helicone

39.0/100 · #4 of 8

Winning prompts (5/5 engines)

No full-coverage prompts this week.

Losing prompts (0/5 engines)

OpenAI proxy with caching and rate limiting
Router that fails over between Claude, GPT, and open-source models

Portkey

36.0/100 · #5 of 8

Winning prompts (5/5 engines)

No full-coverage prompts this week.

Losing prompts (0/5 engines)

LLM observability platform for tracking prompts, costs, and latency
OpenAI proxy with caching and rate limiting
LLM analytics tool for debugging agent workflows

OpenRouter

35.0/100 · #6 of 8

Winning prompts (5/5 engines)

No full-coverage prompts this week.

Losing prompts (0/5 engines)

LLM observability platform for tracking prompts, costs, and latency
OpenAI proxy with caching and rate limiting
LLM analytics tool for debugging agent workflows

Vercel AI Gateway

30.0/100 · #7 of 8

Winning prompts (5/5 engines)

No full-coverage prompts this week.

Losing prompts (0/5 engines)

LLM observability platform for tracking prompts, costs, and latency
LLM analytics tool for debugging agent workflows

Braintrust

29.0/100 · #8 of 8

Winning prompts (5/5 engines)

No full-coverage prompts this week.

Losing prompts (0/5 engines)

Best LLM gateway for routing across multiple model providers
OpenAI proxy with caching and rate limiting
Router that fails over between Claude, GPT, and open-source models