Models

gpt-oss-20b

darkbloom · 131K ctx · 20 providers · updated 2026-08-05 00:15 UTC

Cheapest
$0.028 Official list · Darkbloom
Avg price
$0.095 /1M
providers
20
Tokens (24h)
26B
Volume (24h)
$1.4K est.
(3 × $0.015 + $0.070) ÷ 4 = $0.028/1M — the $/1M figure blends input and output at 3:1 — that is the only way models with two different rates can be ranked on one axis. Your actual rate depends on your own input:output mix — input only $0.015, 3:1 $0.028, 1:1 $0.042, output only $0.070 per 1M.

The board shows this model's cheapest measured offer ($0.028/1M); the smaller figure under it is the mean of the 33 offers listed below ($0.095/1M). Add those up and divide — it matches. 24h token volume (26B) is sourced from OpenRouter public rankings — today's only public source of absolute per-model token counts (Vercel AI Gateway publishes share-of-traffic percentages); as more sources come online, the per-router breakdown will sum to this total here.

ProviderRouter / Source$ in /1M$ out /1MTTFTTPSUptimeEvidence
darkbloom Pricebook (LiteLLM) $0.015$0.070 Official list
openrouter Pricebook (LiteLLM) $0.020$0.100 Official list
CoreWeave (fp4) OpenRouter $0.030$0.130 547ms 18 98.9% API-measured
DeepInfra (bf16) OpenRouter $0.030$0.140 329ms 86 99.7% API-measured
deepinfra HuggingFace Router $0.030$0.140 266ms 86 API-measured
Parasail (fp4) OpenRouter $0.030$0.150 533ms 54 85.9% API-measured
deepinfra Pricebook (LiteLLM) $0.040$0.150 Official list
ovhcloud Pricebook (LiteLLM) $0.040$0.150 Official list
novita Pricebook (LiteLLM) $0.040$0.150 Official list
novita HuggingFace Router $0.040$0.150 385ms 43 API-measured
novita-ai Novita AI $0.040$0.150 API-measured
Novita (fp4) OpenRouter $0.040$0.150 481ms 38 100.0% API-measured
Phala OpenRouter $0.040$0.150 401ms 57 97.6% API-measured
SiliconFlow (fp8) OpenRouter $0.040$0.180 1187ms 35 100% API-measured
ovhcloud HuggingFace Router $0.050$0.180 272ms 82 API-measured
together_ai Pricebook (LiteLLM) $0.050$0.200 Official list
Together OpenRouter $0.050$0.200 229ms 76 82.1% API-measured
nscale HuggingFace Router $0.050$0.200 649ms 146 API-measured
together HuggingFace Router $0.050$0.200 1076ms 74 API-measured
burncloud burncloud $0.050$0.200 API-measured
Amazon Bedrock (eu-west-1) OpenRouter $0.070$0.150 API-measured
Amazon Bedrock OpenRouter $0.070$0.150 353ms 342 99.9% API-measured
Google (us-central1) OpenRouter $0.070$0.250 2506ms 172 87.8% API-measured
tensormesh Pricebook (LiteLLM) $0.070$0.280 Official list
fireworks_ai Pricebook (LiteLLM) $0.070$0.300 Official list
fireworks-ai HuggingFace Router $0.070$0.300 364ms 52 API-measured
requesty Requesty $0.070$0.300 API-measured
Fireworks OpenRouter $0.070$0.300 436ms 47 16.4% API-measured
groq Pricebook (LiteLLM) $0.075$0.300 Official list
Groq OpenRouter $0.075$0.300 300ms 173 99.8% API-measured
replicate Pricebook (LiteLLM) $0.090$0.360 Official list
groq HuggingFace Router $0.100$0.500 277ms 633 API-measured
cloudflare Pricebook (LiteLLM) $0.200$0.300 Official list

Who burns this model

Top apps routing traffic to this model over the last 30 days, via OpenRouter's public stats. Only apps that opted into tracking appear, and only the top few are published — this is not the full demand picture. These figures are a 30-day window and are NOT comparable with the cumulative totals on the Apps board.

AppTokens (30d)Requests
JobLeads LLM 85B 15M
HeyDitto 24B 4M
FVChat 20B 5M
Central Command OpenClaw 19B 1M
Craft 17B 2M

Volume trend

35B 30d d-29 · 19B tokensd-28 · 26B tokensd-27 · 34B tokensd-26 · 25B tokensd-25 · 35B tokensd-24 · 23B tokensd-23 · 0 tokensd-22 · 24B tokensd-21 · 31B tokensd-20 · 32B tokensd-19 · 30B tokensd-18 · 0 tokensd-17 · 0 tokensd-16 · 0 tokensd-15 · 0 tokensd-14 · 0 tokensd-13 · 0 tokensd-12 · 0 tokensd-11 · 0 tokensd-10 · 0 tokensd-9 · 18B tokensd-8 · 0 tokensd-7 · 0 tokensd-6 · 0 tokensd-5 · 0 tokensd-4 · 0 tokensd-3 · 20B tokensd-2 · 26B tokensd-1 · 27B tokensd-0 · 26B tokens

Last 30 days: 19B → 26B (+33%); range 0–35B tokens/day. Source: OpenRouter public rankings; refreshed hourly.

Data & method

Cheapest ranked by blended = (3×in + out) ÷ 4; performance columns show only when a source provides them, never estimated. “Official list” = LiteLLM open pricebook (MIT); “API-measured” = provider/aggregator public APIs. Machine-readable data: /api/v1/models/gpt-oss-20b.json. Free reuse requires attribution to tkx.org.

About

GPT-OSS-20B is a large language model served by multiple inference providers on AI model marketplaces, known for its capabilities in natural language processing tasks. It is listed under the organization tag "darkbloom".

AI-generated summary from public information — is this yours?

FAQ

How much does gpt-oss-20b cost?

Cheapest measured offer right now: $0.0145/1M input, $0.07/1M output via darkbloom — across 20 tracked providers, refreshed hourly.

Who serves gpt-oss-20b?

20 providers currently serve it. The table above lists each with input/output price, TTFT, TPS, uptime and an evidence tier.

How heavily is gpt-oss-20b used?

gpt-oss-20b routed 26B tokens in the last 24h. The daily bars above show routed volume over recent days. Source: OpenRouter public rankings — one of the few routers publishing per-model usage; TKX refreshes hourly.

Related analysis

Back to rankings