All models
OpenAI
gpt-oss-20b fastruns locally
openai/gpt-oss-20b
gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for...
Input / 1M
$0.030
USD per million tokens
Output / 1M
$0.13
USD per million tokens
Est. cost / token
$0.000000055
$0.055 per 1M, 3:1 mix
Context window
131K
max output 118K
Major features
Open weightsCachingReasoningJSON schemaTools
Task suitability (heuristic 1–5, not a benchmark)
- coding3
- reasoning3
- math3
- summarization4
- extraction4
- classification4
- chat3
- creative3
- translation4
- vision1
- long context3
- agentic3
Run it locally
Will it run on my machine? →ollama run gpt-oss:20b| Quantization | Quality | Memory needed (8K ctx) |
|---|---|---|
| Q2_K | poor | 11.4 GB |
| Q3_K_M | fair | 13 GB |
| Q4_K_Mrecommended | good | 15.7 GB |
| Q5_K_M | very good | 17.9 GB |
| Q6_K | excellent | 20.3 GB |
| Q8_0 | near-lossless | 25.6 GB |
| F16 | lossless | 46.3 GB |
Estimates. Mixture-of-experts: memory covers all parameters; only the active ones are used per token, so it runs faster than its size suggests.
Price history
| Recorded | Input | Output | Cached | Source |
|---|---|---|---|---|
| 2026-11-05 03:20 | $0.030 | $0.13 | $0.030 | openrouter |
A row is added whenever a sync detects a price change. Vendor pricing source
Pricing (USD per 1M tokens)
- Input
- $0.030
- Output
- $0.13
- Cached input
- $0.030
- Blended (3:1)
- $0.055
Details
- Company
- OpenAI
- Version
- —
- Family
- GPT
- Released
- 2025-08-05
- Parameters
- 21B (3.6B active)
- Max output
- 118K
- Open weights
- Yes
- Inputs
- text
- Last synced
- 25 min ago