NEWMEN
← back to all models

DeepSeek: DeepSeek V3.1

deepseek/deepseek-chat-v3.1

DeepSeekopen-weight164K context7 providersIntelligence 71.0· est.

Cheapest provider

$0.21 / 1M

DeepInfra

Fastest provider (p95)

158 tok/s

Google

Intelligence (estimated)

71.0

Family + generation + popularity + price tier

Per-provider performance

Latency / throughput / uptime / price measured across providers over the last 30 minutes of live traffic. This is what proves “sourced cheapest” — Atlas mode draws on these per call to serve the cheapest path that holds quality.

ProviderQuantInput $/1MOutput $/1MLatency p50 / p95Throughput p50 / p95Uptime 30mSuccess
DeepInfraq4· fp4$0.2100$0.7900810ms / 2456ms11 / 15 tok/s99.73%85.8%
Novitaq8· fp8$0.2700$1.00001746ms / 4343ms17 / 35 tok/s100.00%100.0%
SiliconFlowq8· fp8$0.2700$1.00001305ms / 5678ms16 / 24 tok/s95.04%95.1%
AtlasCloudq8· fp8$0.3000$0.95001573ms / 3991ms6 / 22 tok/s99.43%99.4%
WandBq8· fp8$0.5500$1.6500349ms / 681ms27 / 42 tok/s99.99%100.0%
Googleundisclosed$0.6000$1.7000694ms / 1520ms55 / 158 tok/s99.39%86.7%
SambaNovaq8· fp8$0.6500$1.50001195ms / 5184ms57 / 155 tok/s98.38%72.8%

“—” means live telemetry hasn’t accumulated enough recent traffic for that endpoint. “undisclosed” means the provider serves the model but doesn’t expose the quantization label (typically running fp8 / int8 internally).

Intelligence estimate

No public benchmark numbers indexed for this model yet, so the leaderboard score is derived from the catalogue data we sync: model family standing, generation, and live usage rank — applied identically to open- and closed-weight models. The heuristic ceiling sits below confirmed-benchmark frontier models so curated rankings stay clearly on top.

Want a curated score? File a benchmark report and we’ll add it to the next sync.

How Atlas mode sources DeepSeek: DeepSeek V3.1

  • Strict mode — pin DeepSeek: DeepSeek V3.1 exactly and we pass it straight through, sourced from the cheapest provider above. The same model, no substitutions — currently DeepInfra at $0.21/1M.
  • Atlas mode — the default. Each call is auto-optimized for the cheapest path that holds quality, at least 5% off going direct from call one and climbing as it ramps. You always see which model served the call and exactly what you saved — thumbs-down anything you don’t like for a full refund.