gpt-oss API pricing
OpenAI's open-weight models, released under Apache 2.0 in August 2025. Many hosts run them; prices here are the cheapest listed on OpenRouter.
gpt-oss-120b
▼ 75% on 13 OctCapability score 140, 28th of 29 current models scored on the Epoch Capabilities Index (range 138–142).
GPT-6 Luna scores higher (157) and costs 21% less on a coding agent: $25.98 vs $32.88 a month.
Source: Epoch Capabilities Index by Epoch AI, CC BY 4.0, read 11 Nov 2026. What the score means
What it costs in practice
gpt-oss-120b at list prices, no batch discount.
| 1,000 chat replies1,500 tokens in and 300 out each | $0.09 |
|---|---|
| Summarising 100 long reports20,000 tokens in and 800 out each | $0.07 |
| A month of a coding agent400 tasks, 85% of input cached | $23.70 |
The coding-agent example assumes 85% of input is served from cache, typical for agents that resend the same context each step. Work out AI agent cost on gpt-oss-120b
Details
gpt-oss-120bCheapest host listed on OpenRouter (Venice). Most other hosts charge $0.03–0.04 / $0.17–0.19.
Checked 13 Oct 2026 against OpenRouter listing .
Estimate your monthly cost
List prices, no batch discount
| Model | Input / output per 1M | Per month |
|---|---|---|
| gpt-oss-120b | $0.030 / $0.15 | $0.60 |
| gpt-oss-20b | $0.018 / $0.090 | $0.36 |
Type amounts like 500K, 25M or 2B.
Price gpt-oss-120b in: Token cost calculator · AI agent cost · Token counter
Cheapest in its class
No current general-purpose workhorse model from another family costs meaningfully less than gpt-oss-120b.
Compare gpt-oss-120b with
Side by side: prices, cached and batch rates, and the cost of the same workload.
Other gpt-oss models
Because the weights are open, you can also run them yourself. See the local LLM calculator.
| Model | Input / 1M | Output / 1M | Cached input | Context | Released |
|---|---|---|---|---|---|
Open weights · via OpenRouter · Source: OpenRouter listing (opens in a new tab) |
$0.018 | $0.090 | — | 128K | 5 Aug 2025 |
gpt-oss-20b details
gpt-oss-20b
$0.018 / $0.090 per 1M tokens128K context
Cheapest host listed on OpenRouter (Darkbloom). DeepInfra charges $0.03 / $0.14.
Checked 13 Oct 2026 against OpenRouter listing .
Compare gpt-oss-120b
No ready-made comparisons for this family yet.
Price history
All changes- The cheapest OpenRouter hosts repriced 6 open-weight models, mostly downwards, including gpt-oss-120b (now $0.03 / $0.15) and Qwen3.8 27B (now $0.039 / $1.25). Source
Every dated price: price records · CSV. Cite as: AI Token Price, gpt-oss pricing, https://aitokenprice.com/pricing/gpt-oss#price-history
OpenAI pricing rules
- Batch and Flex processing cost 50% less.
- For GPT-5.4 and later, prompts longer than 272K tokens are billed at 2× the input price and 1.5× the output price.
- Cached input is billed at a tenth of the input price on most models, and a twentieth on GPT-6.1 Sol.
Questions
How much does gpt-oss-120b cost?
As of 13 Oct 2026, gpt-oss-120b costs $0.03 per million input tokens and $0.15 per million output tokens. Prices checked 13 Oct 2026.
What is the cheapest gpt-oss model?
gpt-oss-20b, at $0.018 / $0.09 per million input / output tokens.
What is gpt-oss-120b's context window?
128K tokens.