OpenAI: gpt-oss-20b
openai/gpt-oss-20bgpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license.
Modalities
In / out per 1M$0.02 / $0.09
Context128K
ReleasedAug 5, 2025
Providers
11 providers serve this model upstream.
| Provider | Input /M | Output /M | Cache read /M | Latency P50 | Throughput P50 | Uptime |
|---|---|---|---|---|---|---|
| $0.03 | $0.14 | $0.03 | 0.52s | 37 tps | 100% | |
 Amazon Bedrockeu-west-1 | $0.07 | $0.15 | Variable | 1.30s | 116 tps | 100% |
| $0.03 | $0.15 | $0.02 | 0.88s | 64 tps | 99.95% | |
| Amazon Bedrock | $0.07 | $0.15 | Variable | 0.41s | 341 tps | 99.95% |
 Darkbloomfp8 | $0.02 | $0.09 | $0.01 | 4.22s | 32 tps | 99.84% |
| $0.02 | $0.10 | Variable | 0.82s | 41 tps | 99.81% | |
 CoreWeavefp4 | $0.03 | $0.13 | $0.03 | 0.41s | 135 tps | 99.81% |
| $0.07 | $0.30 | $0.04 | 0.49s | 297 tps | 99.65% | |
 DeepInfraus · bf16 | $0.03 | $0.14 | Variable | 0.30s | 81 tps | 99.45% |
 Google Vertexus-central1 | $0.07 | $0.25 | Variable | 0.63s | 151 tps | 95.63% |
 SiliconFlowfp8 | $0.04 | $0.18 | Variable | 1.13s | 33 tps | 87.04% |
Market-wide measurements, refreshed every 15 minutes. They describe who serves gpt-oss-20b in general, not how ProjectCOZY routes your requests.
Pricing
List rates per 1M tokens — what gpt-oss-20b costs pay-as-you-go elsewhere, not what you pay here.
Input
$0.02
per 1M tokens
Output
$0.09
per 1M tokens
Cache read
$0.01
per 1M cached tokens
Included in every pass
gpt-oss-20b is included in every ProjectCOZY pass at no extra cost — one flat rate, no per-token metering. The rates above are there for comparison.
Performance
Time to first token and output speed across every provider, over the last 3 days.
0.0s5.0s10.0s15.0s20.0s
Oct 11, 12 AMOct 11, 8 AMOct 11, 4 PMOct 12, 12 AMOct 12, 8 AMOct 12, 4 PMOct 13, 12 AMOct 13, 8 AMOct 13, 4 PMOct 14, 12 AMOct 14, 8 AM
DekaLLM SiliconFlow Groq Darkbloom AkashML Parasail CoreWeave DeepInfra (us) Amazon Bedrock Google Vertex (us-central1) Amazon Bedrock (eu-west-1)
0 tok/s500 tok/s1000 tok/s1500 tok/s2000 tok/s
Oct 11, 12 AMOct 11, 8 AMOct 11, 4 PMOct 12, 12 AMOct 12, 8 AMOct 12, 4 PMOct 13, 12 AMOct 13, 8 AMOct 13, 4 PMOct 14, 12 AMOct 14, 8 AM
DekaLLM SiliconFlow Groq Darkbloom AkashML Parasail CoreWeave DeepInfra (us) Amazon Bedrock Google Vertex (us-central1) Amazon Bedrock (eu-west-1)
Uptime
100% of requests were served successfully over the last 3 days, across all providers.
Last 3 daysAvailability 98.02%
Oct 11, 2:00 PMNow
Availability over the last 24 hoursAvailability 98.56%
100% 97% 93% 90%
Oct 13, 2:40 PMNow
Apps
Tools that most often run gpt-oss-20b, by tokens this week.
Activity
989.3B tokens processed market-wide in the last month.
020.0B40.0B60.0B80.0B
Sep 14Sep 19Sep 24Sep 29Oct 4Oct 9Oct 14
Tokens
FAQ
Which model id do I use for gpt-oss-20b?
openai/gpt-oss-20b. Send it as the model field from any OpenAI-compatible client, against http://192.168.1.55/v1.
How much context does gpt-oss-20b hold?
131,072 tokens per request, of which up to 32,768 can be the response.
What can it accept and return?
Input: text. Output: text.
What does gpt-oss-20b cost on a pass?
Nothing beyond the pass itself. Pay-as-you-go elsewhere it lists at $0.02 per 1M input tokens and $0.09 per 1M output tokens.
How current is its knowledge?
Training data runs to Jun 2024. Anything later has to come from your prompt or a tool call.
Who built gpt-oss-20b?
OpenAI, in the gpt family.
Explore
Other models your pass can call.
