ninjachat api · benchmarks

Live AI model benchmarks.

Which provider is fastest for a given model? Which is most reliable? We transact text, images and video across every provider — so we can measure it. These leaderboards are computed from real API request traffic, not vendor marketing.

7-day window · as of Thu, 19 Nov 2026 08:40:25 UTC

Provider performance

Speed and reliability, by provider.

Every measured provider across all models it served in the last 7 days. Success counts HTTP 5xx as failures; client errors (4xx) and rate limits are excluded, so one bad caller can't sink a score.

ProviderSuccessTypical p50Typical TTFT
Anthropic100.0%6.40 s—
BytePlus100.0%2.64 s—
DeepInfra100.0%11.31 s8.49 s
Fireworks100.0%6.11 s2.35 s
Google99.2%9.59 s2.46 s
Novita AI100.0%9.62 s7.40 s
OpenAI100.0%5.77 s1.36 s
OpenRouter100.0%23.91 s21.48 s
Replicate94.0%10.68 s—
xAI100.0%13.88 s12.79 s
By model

Fastest provider, per model.

For each model, the provider currently serving it fastest. Open a model for the full provider comparison and its latency, uptime and price history.

Price / performance

Best value.

Output tokens per second per cent of typical request price — measured throughput against what you actually pay. Chat models with measured performance.

ModelTokens/secPriceValue (tok/s per ¢)
DeepSeek V4 Flash130.4$0.00101330.6
Step 3.7 Flash217.8$0.00221013
MiMo V2.572.3$0.0010737.8
Gemini 2.5 Flash126.9$0.0040317.3
GPT-4.1 Mini111.7$0.0036310.3
GPT-5 Mini85.7$0.0033263.7
Gemini 3 Flash39.3$0.005571.5
DeepSeek V4 Pro70.9$0.01258.2
Gemini 3.5 Flash95$0.01757.6
Gemini 3.7 Flash37$0.007549.3

Measured from live API traffic. GET /api/v1/health · Status