GPT-OSS 120B vs Gemini 2.5 Flash
GPT-OSS 120B is an open-weight chat model you can call on this endpoint; Gemini 2.5 Flash is a closed model from Google. Below: what each costs for three realistic workloads, the limits that apply, and when to choose which.
Google figures are public list prices and documented limits, checked 2026-09-16 against the provider page. Inference APIs prices are read live from the price list.
- Cost: GPT-OSS 120B is cheaper in all three workloads, by 45–67%.
- Limits: Google: Free tier has per-model daily request quotas and returns 429 RESOURCE_EXHAUSTED when they are used; paid tier is self-serve. Here: no per-minute or per-day caps.
Side by side
| GPT-OSS 120B | Gemini 2.5 Flash | |
|---|---|---|
| Served by | Inference APIs | |
| Model author | OpenAI | |
| Model id | openai/gpt-oss-120b | gemini-2.5-flash |
| Input price | $0.20 / 1M tokens | $0.30 / 1M tokens |
| Output price | $0.80 / 1M tokens | $2.50 / 1M tokens |
| Context window | 131K tokens | 1M tokens |
| Capabilities | Chat, Reasoning, Tool calling, JSON mode | Chat, Reasoning, Vision, Tool calling, JSON mode |
| Weights | Open | Closed |
| Status | Available | Available |
| Limits | No per-minute or per-day request or token caps; usage draws on a prepaid balance | Free tier has per-model daily request quotas and returns 429 RESOURCE_EXHAUSTED when they are used; paid tier is self-serve. |
| Time to first token | 0.46 s | not measured by us |
| Output speed | 45.8 tokens / s | not measured by us |
What three workloads cost
List prices applied to the same work. The cheaper side of each row is in bold.
| Workload | GPT-OSS 120B | Gemini 2.5 Flash | Difference |
|---|---|---|---|
| Chat assistant — 10,000 turns of 800 tokens in, 300 out | $4.00 | $9.90 | 60% |
| Long prompts or RAG — 10,000 requests of 8,000 in, 500 out | $20.00 | $36.50 | 45% |
| Generation-heavy — 10,000 requests of 500 in, 2,000 out | $17.00 | $51.50 | 67% |
GPT-OSS 120B is cheaper in all three workloads, by 45–67%. Reasoning models bill their thinking as output tokens, so real output counts run higher than the visible answer; treat the generation-heavy row as a floor for them.
When to choose which
- Cost is the deciding factor; it is cheaper in every workload above
- You want open weights: the same model can be moved to another host or self-hosted later
- You are replacing llama-3.3-70b-versatile and want the model Groq itself recommends
- You build agents: tool calling, JSON-schema output and a visible reasoning trace are all verified on this endpoint
- Your workload is output-heavy; output tokens are the expensive side of its price
- You need more than 131K tokens of context (it takes 1M)
- You need to send images; GPT-OSS 120B is text-only
- You need multimodal input (images, audio, video)
- A free tier for development matters to you
Price, context and speed are measurable; answer quality on your task is not something a table can settle. Both take the same request, so the honest test is to run your own prompts through each.
Measured speed
Medians of three streamed runs on 2026-09-17, public endpoint, a prompt that asks for about 120 words. Time to first token is the first token of any kind, reasoning included; output speed counts every generated token from that point. Reasoning models then think before the visible answer starts: GPT-OSS 120B began answering after 1.38 s at default settings, which you can shorten with the reasoning controls on each model page. Expect ±30% with time of day. We did not measure Google and do not quote other people's numbers; read "not measured" as unknown, not slow. Method and raw numbers: speed measurements.
Switching from Google
Both speak the OpenAI wire format. The edit is the base URL, the key and the model id; then re-run your own prompts, because it is a different model.
import os
from openai import OpenAI
# before: Google
# client = OpenAI(base_url="https://generativelanguage.googleapis.com/v1beta/openai/", api_key=os.environ["GOOGLE_API_KEY"])
# MODEL = "gemini-2.5-flash"
# after: Inference APIs
client = OpenAI(base_url="https://api.inferenceapis.com/v1", api_key=os.environ["INFERENCE_API_KEY"])
MODEL = "openai/gpt-oss-120b"
resp = client.chat.completions.create(model=MODEL, messages=[{"role": "user", "content": "Hello"}], max_tokens=500)More detail, including Node, LangChain and LiteLLM: switching OpenAI-compatible providers.
FAQ
Which is cheaper, GPT-OSS 120B or Gemini 2.5 Flash?
GPT-OSS 120B is cheaper in all three workloads, by 45–67%. The table above uses list prices: GPT-OSS 120B $0.20 / 1M tokens input, $0.80 / 1M tokens output; Gemini 2.5 Flash $0.30 / 1M tokens input, $2.50 / 1M tokens output.
Can I switch from Gemini 2.5 Flash to GPT-OSS 120B without rewriting code?
Yes, if you call it through an OpenAI-compatible client. Change the base URL to https://api.inferenceapis.com/v1, swap the API key, and set the model to openai/gpt-oss-120b. Because it is a different model, re-run your prompts and evaluations before moving production traffic.
What limits apply to Gemini 2.5 Flash?
Free tier has per-model daily request quotas and returns 429 RESOURCE_EXHAUSTED when they are used; paid tier is self-serve. On Inference APIs there are no per-minute or per-day request or token caps; usage draws on a prepaid balance.
Why does GPT-OSS 120B use more output tokens than the visible answer?
It is a reasoning model: it thinks before it answers and the thinking is billed as output tokens. Budget for that in output-heavy workloads, and set max_tokens to a few hundred or more so the answer is not cut off.
Spotted a price or limit that has changed? Tell us and we will re-check the provider page.