{
"service": "InferenceLatency.com",
"endpoint": "throughput",
"description": "Combined latency and throughput benchmarking",
"providers": [
{
"provider": "OpenAI",
"model": "GPT-4o",
"metrics": {
"latency_ms": 1095,
"throughput_tokens_per_sec": 15.53
},
"tokens_generated": 17,
"elapsed_seconds": 1.095,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.0025,
"estimated_cost_usd": 4.3e-05
},
"model_metadata": {
"release_date": "2024-05-13",
"context_length": 128000,
"hardware": "GPU",
"api_type": "OpenAI-compatible"
},
"infra": {
"schema_url": "https://platform.openai.com/docs/api-reference",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://platform.openai.com/docs",
"api_docs": "https://platform.openai.com/docs/api-reference/chat"
},
"test_metadata": {
"request_hash": "dc2b6d757486",
"trace_id": "trace-e1a9b6b4"
},
"history": {
"latency_trend_7d": [
752,
750,
748,
749,
751,
750,
1095
],
"throughput_trend_7d": [
31.9,
32.1,
32.0,
32.2,
32.0,
32.0,
15.53
]
}
},
{
"provider": "Groq",
"model": "Llama3-8B",
"metrics": {
"latency_ms": null,
"throughput_tokens_per_sec": null
},
"error": "Client error '400 Bad Request' for url 'https://api.groq.com/openai/v1/chat/completions'\nFor more information check: https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/400",
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 0.0,
"error_rate_percent": 100.0,
"status": "unhealthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.00027,
"estimated_cost_usd": 0.0
},
"model_metadata": {
"release_date": "2024-04-18",
"context_length": 8192,
"hardware": "LPU",
"api_type": "OpenAI-compatible",
"schema_url": "https://console.groq.com/docs/openai",
"human_readable_url": "https://console.groq.com/docs",
"api_docs": "https://console.groq.com/docs/api-reference"
},
"infra": {
"schema_url": "https://console.groq.com/docs/openai",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://console.groq.com/docs",
"api_docs": "https://console.groq.com/docs/api-reference"
},
"test_metadata": {
"request_hash": "da1ed9b456ed",
"trace_id": "trace-7f20559c"
},
"history": {
"latency_trend_7d": [
952,
950,
948,
949,
951,
950,
950
],
"throughput_trend_7d": [
121.9,
122.1,
122.0,
122.2,
122.0,
122.0,
122.01
]
}
},
{
"provider": "Claude",
"model": "Claude Sonnet 4",
"metrics": {
"latency_ms": 2471,
"throughput_tokens_per_sec": 12.55
},
"tokens_generated": 31,
"elapsed_seconds": 2.471,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.003,
"estimated_cost_usd": 9.3e-05
},
"model_metadata": {
"release_date": "2024-10-22",
"context_length": 200000,
"hardware": "GPU",
"api_type": "Anthropic-native"
},
"infra": {
"schema_url": "https://docs.anthropic.com/en/api/messages",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://docs.anthropic.com",
"api_docs": "https://docs.anthropic.com/en/api"
},
"test_metadata": {
"request_hash": "bb1a07893be0",
"trace_id": "trace-b646eda2"
},
"history": {
"latency_trend_7d": [
1152,
1150,
1148,
1149,
1151,
1150,
2471
],
"throughput_trend_7d": [
21.9,
22.1,
22.0,
22.2,
22.0,
22.0,
12.55
]
}
},
{
"provider": "OpenRouter",
"model": "Mistral",
"metrics": {
"latency_ms": 1401,
"throughput_tokens_per_sec": 16.42
},
"tokens_generated": 23,
"elapsed_seconds": 1.401,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.00018,
"estimated_cost_usd": 4e-06
},
"model_metadata": {
"release_date": "2024-02-26",
"context_length": 32768,
"hardware": "GPU",
"api_type": "OpenAI-compatible"
},
"infra": {
"schema_url": "https://openrouter.ai/docs",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://openrouter.ai/docs",
"api_docs": "https://openrouter.ai/docs/api"
},
"test_metadata": {
"request_hash": "6e0c7b85890a",
"trace_id": "trace-f5d01234"
},
"history": {
"latency_trend_7d": [
652,
650,
648,
649,
651,
650,
1401
],
"throughput_trend_7d": [
41.9,
42.1,
42.0,
42.2,
42.0,
42.0,
16.42
]
}
},
{
"provider": "Google Gemini",
"model": "Gemini-2.0-Flash",
"metrics": {
"latency_ms": 583,
"throughput_tokens_per_sec": 18.88
},
"tokens_generated": 11,
"elapsed_seconds": 0.583,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.00075,
"estimated_cost_usd": 8e-06
},
"model_metadata": {
"release_date": "2024-12-11",
"context_length": 2000000,
"hardware": "TPU",
"api_type": "Google-native"
},
"infra": {
"schema_url": "https://ai.google.dev/api",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://ai.google.dev/docs",
"api_docs": "https://ai.google.dev/api/generate-content"
},
"test_metadata": {
"request_hash": "10d1eaa1ac9d",
"trace_id": "trace-1ecf93d9"
},
"history": {
"latency_trend_7d": [
352,
350,
348,
349,
351,
350,
583
],
"throughput_trend_7d": [
91.9,
92.1,
92.0,
92.2,
92.0,
92.0,
18.88
]
}
},
{
"provider": "Together AI",
"model": "Llama3.1-8B-Turbo",
"metrics": {
"latency_ms": 361,
"throughput_tokens_per_sec": 88.65
},
"tokens_generated": 32,
"elapsed_seconds": 0.361,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.0002,
"estimated_cost_usd": 6e-06
},
"model_metadata": {
"release_date": "2024-01-15",
"context_length": 32768,
"hardware": "GPU",
"api_type": "OpenAI-compatible"
},
"infra": {
"schema_url": "https://docs.together.ai/docs/inference-models",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://docs.together.ai",
"api_docs": "https://docs.together.ai/reference/chat-completions"
},
"test_metadata": {
"request_hash": "c35de0059f8b",
"trace_id": "trace-4fa3ad56"
},
"history": {
"latency_trend_7d": [
120,
118,
116,
117,
119,
118,
361
],
"throughput_trend_7d": [
145.9,
146.1,
146.0,
146.2,
146.0,
146.0,
88.65
]
}
},
{
"provider": "Fireworks AI",
"model": "Llama3.1-8B",
"metrics": {
"latency_ms": 440,
"throughput_tokens_per_sec": 72.72
},
"tokens_generated": 32,
"elapsed_seconds": 0.44,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.0002,
"estimated_cost_usd": 6e-06
},
"model_metadata": {
"release_date": "2024-04-15",
"context_length": 8192,
"hardware": "GPU",
"api_type": "OpenAI-compatible"
},
"infra": {
"schema_url": "https://readme.fireworks.ai/reference/createchatcompletion",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://fireworks.ai/",
"api_docs": "https://readme.fireworks.ai/docs"
},
"test_metadata": {
"request_hash": "f8bbfc248c7e",
"trace_id": "trace-5a6f10ee"
},
"history": {
"latency_trend_7d": [
180,
178,
176,
177,
179,
178,
440
],
"throughput_trend_7d": [
135.9,
136.1,
136.0,
136.2,
136.0,
136.0,
72.72
]
}
},
{
"provider": "HF GPT OSS 120B (Cerebras)",
"model": "GPT OSS 120B",
"metrics": {
"latency_ms": 412,
"throughput_tokens_per_sec": 121.32
},
"tokens_generated": 50,
"elapsed_seconds": 0.412,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.0008,
"estimated_cost_usd": 4e-05
},
"model_metadata": {
"release_date": "2025-01-07",
"context_length": 128000,
"hardware": "Cerebras WSE",
"api_type": "OpenAI-compatible"
}
}
],
"rankings": {
"by_latency": [
{
"provider": "Together AI",
"model": "Llama3.1-8B-Turbo",
"metrics": {
"latency_ms": 361,
"throughput_tokens_per_sec": 88.65
},
"tokens_generated": 32,
"elapsed_seconds": 0.361,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.0002,
"estimated_cost_usd": 6e-06
},
"model_metadata": {
"release_date": "2024-01-15",
"context_length": 32768,
"hardware": "GPU",
"api_type": "OpenAI-compatible"
},
"infra": {
"schema_url": "https://docs.together.ai/docs/inference-models",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://docs.together.ai",
"api_docs": "https://docs.together.ai/reference/chat-completions"
},
"test_metadata": {
"request_hash": "c35de0059f8b",
"trace_id": "trace-4fa3ad56"
},
"history": {
"latency_trend_7d": [
120,
118,
116,
117,
119,
118,
361
],
"throughput_trend_7d": [
145.9,
146.1,
146.0,
146.2,
146.0,
146.0,
88.65
]
}
},
{
"provider": "HF GPT OSS 120B (Cerebras)",
"model": "GPT OSS 120B",
"metrics": {
"latency_ms": 412,
"throughput_tokens_per_sec": 121.32
},
"tokens_generated": 50,
"elapsed_seconds": 0.412,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.0008,
"estimated_cost_usd": 4e-05
},
"model_metadata": {
"release_date": "2025-01-07",
"context_length": 128000,
"hardware": "Cerebras WSE",
"api_type": "OpenAI-compatible"
}
},
{
"provider": "Fireworks AI",
"model": "Llama3.1-8B",
"metrics": {
"latency_ms": 440,
"throughput_tokens_per_sec": 72.72
},
"tokens_generated": 32,
"elapsed_seconds": 0.44,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.0002,
"estimated_cost_usd": 6e-06
},
"model_metadata": {
"release_date": "2024-04-15",
"context_length": 8192,
"hardware": "GPU",
"api_type": "OpenAI-compatible"
},
"infra": {
"schema_url": "https://readme.fireworks.ai/reference/createchatcompletion",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://fireworks.ai/",
"api_docs": "https://readme.fireworks.ai/docs"
},
"test_metadata": {
"request_hash": "f8bbfc248c7e",
"trace_id": "trace-5a6f10ee"
},
"history": {
"latency_trend_7d": [
180,
178,
176,
177,
179,
178,
440
],
"throughput_trend_7d": [
135.9,
136.1,
136.0,
136.2,
136.0,
136.0,
72.72
]
}
},
{
"provider": "Google Gemini",
"model": "Gemini-2.0-Flash",
"metrics": {
"latency_ms": 583,
"throughput_tokens_per_sec": 18.88
},
"tokens_generated": 11,
"elapsed_seconds": 0.583,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.00075,
"estimated_cost_usd": 8e-06
},
"model_metadata": {
"release_date": "2024-12-11",
"context_length": 2000000,
"hardware": "TPU",
"api_type": "Google-native"
},
"infra": {
"schema_url": "https://ai.google.dev/api",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://ai.google.dev/docs",
"api_docs": "https://ai.google.dev/api/generate-content"
},
"test_metadata": {
"request_hash": "10d1eaa1ac9d",
"trace_id": "trace-1ecf93d9"
},
"history": {
"latency_trend_7d": [
352,
350,
348,
349,
351,
350,
583
],
"throughput_trend_7d": [
91.9,
92.1,
92.0,
92.2,
92.0,
92.0,
18.88
]
}
},
{
"provider": "OpenAI",
"model": "GPT-4o",
"metrics": {
"latency_ms": 1095,
"throughput_tokens_per_sec": 15.53
},
"tokens_generated": 17,
"elapsed_seconds": 1.095,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.0025,
"estimated_cost_usd": 4.3e-05
},
"model_metadata": {
"release_date": "2024-05-13",
"context_length": 128000,
"hardware": "GPU",
"api_type": "OpenAI-compatible"
},
"infra": {
"schema_url": "https://platform.openai.com/docs/api-reference",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://platform.openai.com/docs",
"api_docs": "https://platform.openai.com/docs/api-reference/chat"
},
"test_metadata": {
"request_hash": "dc2b6d757486",
"trace_id": "trace-e1a9b6b4"
},
"history": {
"latency_trend_7d": [
752,
750,
748,
749,
751,
750,
1095
],
"throughput_trend_7d": [
31.9,
32.1,
32.0,
32.2,
32.0,
32.0,
15.53
]
}
},
{
"provider": "OpenRouter",
"model": "Mistral",
"metrics": {
"latency_ms": 1401,
"throughput_tokens_per_sec": 16.42
},
"tokens_generated": 23,
"elapsed_seconds": 1.401,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.00018,
"estimated_cost_usd": 4e-06
},
"model_metadata": {
"release_date": "2024-02-26",
"context_length": 32768,
"hardware": "GPU",
"api_type": "OpenAI-compatible"
},
"infra": {
"schema_url": "https://openrouter.ai/docs",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://openrouter.ai/docs",
"api_docs": "https://openrouter.ai/docs/api"
},
"test_metadata": {
"request_hash": "6e0c7b85890a",
"trace_id": "trace-f5d01234"
},
"history": {
"latency_trend_7d": [
652,
650,
648,
649,
651,
650,
1401
],
"throughput_trend_7d": [
41.9,
42.1,
42.0,
42.2,
42.0,
42.0,
16.42
]
}
},
{
"provider": "Claude",
"model": "Claude Sonnet 4",
"metrics": {
"latency_ms": 2471,
"throughput_tokens_per_sec": 12.55
},
"tokens_generated": 31,
"elapsed_seconds": 2.471,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.003,
"estimated_cost_usd": 9.3e-05
},
"model_metadata": {
"release_date": "2024-10-22",
"context_length": 200000,
"hardware": "GPU",
"api_type": "Anthropic-native"
},
"infra": {
"schema_url": "https://docs.anthropic.com/en/api/messages",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://docs.anthropic.com",
"api_docs": "https://docs.anthropic.com/en/api"
},
"test_metadata": {
"request_hash": "bb1a07893be0",
"trace_id": "trace-b646eda2"
},
"history": {
"latency_trend_7d": [
1152,
1150,
1148,
1149,
1151,
1150,
2471
],
"throughput_trend_7d": [
21.9,
22.1,
22.0,
22.2,
22.0,
22.0,
12.55
]
}
}
],
"by_throughput": [
{
"provider": "HF GPT OSS 120B (Cerebras)",
"model": "GPT OSS 120B",
"metrics": {
"latency_ms": 412,
"throughput_tokens_per_sec": 121.32
},
"tokens_generated": 50,
"elapsed_seconds": 0.412,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.0008,
"estimated_cost_usd": 4e-05
},
"model_metadata": {
"release_date": "2025-01-07",
"context_length": 128000,
"hardware": "Cerebras WSE",
"api_type": "OpenAI-compatible"
}
},
{
"provider": "Together AI",
"model": "Llama3.1-8B-Turbo",
"metrics": {
"latency_ms": 361,
"throughput_tokens_per_sec": 88.65
},
"tokens_generated": 32,
"elapsed_seconds": 0.361,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.0002,
"estimated_cost_usd": 6e-06
},
"model_metadata": {
"release_date": "2024-01-15",
"context_length": 32768,
"hardware": "GPU",
"api_type": "OpenAI-compatible"
},
"infra": {
"schema_url": "https://docs.together.ai/docs/inference-models",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://docs.together.ai",
"api_docs": "https://docs.together.ai/reference/chat-completions"
},
"test_metadata": {
"request_hash": "c35de0059f8b",
"trace_id": "trace-4fa3ad56"
},
"history": {
"latency_trend_7d": [
120,
118,
116,
117,
119,
118,
361
],
"throughput_trend_7d": [
145.9,
146.1,
146.0,
146.2,
146.0,
146.0,
88.65
]
}
},
{
"provider": "Fireworks AI",
"model": "Llama3.1-8B",
"metrics": {
"latency_ms": 440,
"throughput_tokens_per_sec": 72.72
},
"tokens_generated": 32,
"elapsed_seconds": 0.44,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.0002,
"estimated_cost_usd": 6e-06
},
"model_metadata": {
"release_date": "2024-04-15",
"context_length": 8192,
"hardware": "GPU",
"api_type": "OpenAI-compatible"
},
"infra": {
"schema_url": "https://readme.fireworks.ai/reference/createchatcompletion",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://fireworks.ai/",
"api_docs": "https://readme.fireworks.ai/docs"
},
"test_metadata": {
"request_hash": "f8bbfc248c7e",
"trace_id": "trace-5a6f10ee"
},
"history": {
"latency_trend_7d": [
180,
178,
176,
177,
179,
178,
440
],
"throughput_trend_7d": [
135.9,
136.1,
136.0,
136.2,
136.0,
136.0,
72.72
]
}
},
{
"provider": "Google Gemini",
"model": "Gemini-2.0-Flash",
"metrics": {
"latency_ms": 583,
"throughput_tokens_per_sec": 18.88
},
"tokens_generated": 11,
"elapsed_seconds": 0.583,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.00075,
"estimated_cost_usd": 8e-06
},
"model_metadata": {
"release_date": "2024-12-11",
"context_length": 2000000,
"hardware": "TPU",
"api_type": "Google-native"
},
"infra": {
"schema_url": "https://ai.google.dev/api",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://ai.google.dev/docs",
"api_docs": "https://ai.google.dev/api/generate-content"
},
"test_metadata": {
"request_hash": "10d1eaa1ac9d",
"trace_id": "trace-1ecf93d9"
},
"history": {
"latency_trend_7d": [
352,
350,
348,
349,
351,
350,
583
],
"throughput_trend_7d": [
91.9,
92.1,
92.0,
92.2,
92.0,
92.0,
18.88
]
}
},
{
"provider": "OpenRouter",
"model": "Mistral",
"metrics": {
"latency_ms": 1401,
"throughput_tokens_per_sec": 16.42
},
"tokens_generated": 23,
"elapsed_seconds": 1.401,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.00018,
"estimated_cost_usd": 4e-06
},
"model_metadata": {
"release_date": "2024-02-26",
"context_length": 32768,
"hardware": "GPU",
"api_type": "OpenAI-compatible"
},
"infra": {
"schema_url": "https://openrouter.ai/docs",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://openrouter.ai/docs",
"api_docs": "https://openrouter.ai/docs/api"
},
"test_metadata": {
"request_hash": "6e0c7b85890a",
"trace_id": "trace-f5d01234"
},
"history": {
"latency_trend_7d": [
652,
650,
648,
649,
651,
650,
1401
],
"throughput_trend_7d": [
41.9,
42.1,
42.0,
42.2,
42.0,
42.0,
16.42
]
}
},
{
"provider": "OpenAI",
"model": "GPT-4o",
"metrics": {
"latency_ms": 1095,
"throughput_tokens_per_sec": 15.53
},
"tokens_generated": 17,
"elapsed_seconds": 1.095,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.0025,
"estimated_cost_usd": 4.3e-05
},
"model_metadata": {
"release_date": "2024-05-13",
"context_length": 128000,
"hardware": "GPU",
"api_type": "OpenAI-compatible"
},
"infra": {
"schema_url": "https://platform.openai.com/docs/api-reference",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://platform.openai.com/docs",
"api_docs": "https://platform.openai.com/docs/api-reference/chat"
},
"test_metadata": {
"request_hash": "dc2b6d757486",
"trace_id": "trace-e1a9b6b4"
},
"history": {
"latency_trend_7d": [
752,
750,
748,
749,
751,
750,
1095
],
"throughput_trend_7d": [
31.9,
32.1,
32.0,
32.2,
32.0,
32.0,
15.53
]
}
},
{
"provider": "Claude",
"model": "Claude Sonnet 4",
"metrics": {
"latency_ms": 2471,
"throughput_tokens_per_sec": 12.55
},
"tokens_generated": 31,
"elapsed_seconds": 2.471,
"health": {
"cold_start_latency_ms": null,
"warm_start_latency_ms": 750,
"availability_percent": 100.0,
"error_rate_percent": 0.0,
"status": "healthy",
"is_cold_start": false
},
"cost_estimate": {
"cost_per_1k_tokens_usd": 0.003,
"estimated_cost_usd": 9.3e-05
},
"model_metadata": {
"release_date": "2024-10-22",
"context_length": 200000,
"hardware": "GPU",
"api_type": "Anthropic-native"
},
"infra": {
"schema_url": "https://docs.anthropic.com/en/api/messages",
"plugin_manifest": "/.well-known/ai-plugin.json",
"human_readable_url": "https://docs.anthropic.com",
"api_docs": "https://docs.anthropic.com/en/api"
},
"test_metadata": {
"request_hash": "bb1a07893be0",
"trace_id": "trace-b646eda2"
},
"history": {
"latency_trend_7d": [
1152,
1150,
1148,
1149,
1151,
1150,
2471
],
"throughput_trend_7d": [
21.9,
22.1,
22.0,
22.2,
22.0,
22.0,
12.55
]
}
}
]
},
"fastest_latency": "Together AI",
"highest_throughput": "HF GPT OSS 120B (Cerebras)",
"total_tested": 8,
"successful_tests": 7,
"failed_tests": 1,
"performance_summary": {
"best_latency_ms": 361,
"best_throughput_tokens_per_sec": 121.32,
"avg_latency_ms": 966,
"avg_throughput_tokens_per_sec": 49.44
},
"ai_guidance": {
"best_for_speed": "Together AI",
"best_for_throughput": "HF GPT OSS 120B (Cerebras)",
"recommendation": "Use Together AI for lowest latency, HF GPT OSS 120B (Cerebras) for highest throughput",
"use_case_guidance": {
"real_time_chat": "Recommended: Together AI (lowest latency)",
"bulk_generation": "Recommended: HF GPT OSS 120B (Cerebras) (highest throughput)",
"balanced_workload": "Consider both Together AI and HF GPT OSS 120B (Cerebras)"
}
},
"human_readable_summary": "\u26a1 Fastest: Together AI (361ms - Good) | \ud83d\ude80 Highest throughput: HF GPT OSS 120B (Cerebras) (121.3 tokens/sec - High) | \u2705 7/8 providers responding",
"timestamp": "2025-09-07T23:20:18.229087Z"
}