OpenAI: gpt-oss-20b
openai/gpt-oss-20b
Cheapest provider
$0.03 / 1M
DeepInfra
Fastest provider (p95)
2464 tok/s
Groq
Intelligence (estimated)
62.0
Family + generation + popularity + price tier
Per-provider performance
Latency / throughput / uptime measured across providers over the last 30 minutes of live traffic. Atlas’s router weighs these per call (with the eval-gate signal) when picking a variant for Standard and Batch tiers.
| Provider | Quant | Input $/1M | Output $/1M | Latency p50 / p95 | Throughput p50 / p95 | Uptime 30m | Success |
|---|---|---|---|---|---|---|---|
| DeepInfra | full· bf16 | $0.0300 | $0.1400 | 413ms / 10679ms | 27 / 30 tok/s | 99.99% | 100.0% |
| Novita | q4· fp4 | $0.0400 | $0.1500 | 616ms / 4099ms | 89 / 110 tok/s | 99.96% | 99.9% |
| Parasail | q4· fp4 | $0.0400 | $0.2000 | 409ms / 3563ms | 117 / 159 tok/s | 99.99% | 99.9% |
| SiliconFlow | q8· fp8 | $0.0400 | $0.1800 | 1933ms / 7760ms | 34 / 95.4 tok/s | 99.93% | 95.6% |
| Together | full· unknown | $0.0500 | $0.2000 | 526ms / 3902ms | 117 / 190 tok/s | 100.00% | 100.0% |
| WandB | q4· fp4 | $0.0500 | $0.2000 | 237ms / 980ms | 282 / 440 tok/s | 100.00% | 100.0% |
| Amazon Bedrock | full· unknown | $0.0700 | $0.1500 | 1383ms / 5660ms | 100 / 235 tok/s | 44.03% | 97.8% |
| Amazon Bedrock | full· unknown | $0.0700 | $0.1500 | 1383ms / 5660ms | 100 / 235 tok/s | 99.47% | 97.8% |
| Fireworks | full· unknown | $0.0700 | $0.3000 | 802ms / 8338ms | 50 / 84 tok/s | 94.54% | 94.0% |
| full· unknown | $0.0700 | $0.2500 | 526ms / 1679ms | 22 / 87 tok/s | 99.91% | 99.9% | |
| Groq | full· unknown | $0.0750 | $0.3000 | 166ms / 1845ms | 447 / 2463.7 tok/s | 99.95% | 99.6% |
| NextBit | q8· fp8 | $0.1000 | $0.4500 | 686ms / 2610ms | 115 / 202 tok/s | 99.37% | 100.0% |
“—” 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.
- ·openai legacy family base
Want a curated score? File a benchmark report and we’ll add it to the next sync.
How Atlas routes OpenAI: gpt-oss-20b
- Realtime tier — direct passthrough at the upstream’s native precision. Best for hard latency / quality guarantees.
- Standard tier — Atlas picks the cheapest provider variant whose quantization has stayed green on your operation’s eval gates. For OpenAI: gpt-oss-20b that’s currently DeepInfra at $0.03/1M.
- Batch tier — async, biggest discount. Roadmapped to use provider batch APIs (OpenAI / Anthropic) where available and queued spot capacity for open-weight workloads.