openai Open weights Reasoning

gpt-oss-120b pricing & benchmarks

Released Aug 5, 2025 · served by 17 providers

Input / 1M tokens$0.030
Output / 1M tokens$0.170
Context window131K
Cached input
Intelligence index23.8
Coding index30.4
Agentic index13.2
1M in + 300K out$0.08

What gpt-oss-120b is

gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases. It activates 5.1B parameters per forward pass and is optimized...

Providers and prices for gpt-oss-120b

The same model costs different amounts depending on who serves it. Prices are per 1M tokens, cheapest first. Uptime and throughput are OpenRouter's rolling measurements, not vendor claims.

ProviderInputOutputContext QuantUptime 24hThroughput
CoreWeavecheapest $0.030 $0.170 131K fp4 96.8%
Novita $0.050 $0.250 131K fp4 89.7%
Google $0.090 $0.360 131K 99.2%
SiliconFlow $0.050 $0.450 131K fp8 84.0%
DigitalOcean $0.070 $0.490 128K 99.8%
Mancer 2 $0.055 $0.500 131K fp8 93.3%
BaseTen $0.100 $0.500 128K fp4 99.8%
DeepInfra $0.150 $0.600 131K bf16 99.8%
Together $0.150 $0.600 131K 91.9%
Nebius $0.150 $0.600 131K fp4 98.4%
Amazon Bedrock $0.150 $0.600 131K
Phala $0.150 $0.600 131K 99.4%
Groq $0.150 $0.600 131K 100.0%
Parasail $0.100 $0.750 131K fp4 89.4%
Mara $0.150 $0.750 131K 86.7%
Cerebras $0.350 $0.750 131K fp16 100.0%
SambaNova $0.140 $0.950 131K 97.2%

gpt-oss-120b benchmark results

Design Arena head-to-head results, as reported through the OpenRouter model API.

CategoryArenaRankEloWin rate
gamedev models #92 1049 40.6%
dataviz models #94 1029 45.1%
3d models #98 958 29.4%
uicomponent models #99 961 35.5%
codecategories models #104 994 33.4%
website models #107 993 32.5%

Cheaper models in the same class

Models scoring within 4 points of gpt-oss-120b on the Intelligence Index, but with a lower output price.

FAQ

How much does the gpt-oss-120b API cost?

$0.030 per 1M input tokens and $0.170 per 1M output tokens. A typical workload of 1M input + 300K output tokens costs about $0.08.

Is gpt-oss-120b free?

No. It is a paid model starting at $0.030 per 1M input tokens, though some providers offer trial credits.

Which provider is cheapest for gpt-oss-120b?

CoreWeave at $0.030 input / $0.170 output per 1M tokens (fp4 quantization).

Can I self-host gpt-oss-120b?

Yes — weights are published as openai/gpt-oss-120b on Hugging Face, so you can run it on your own hardware.