About
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...
Related Models
Pricing
Input
$0.03
per 1M tokens
Output
$0.15
per 1M tokens
Blended
$0.06
per 1M tokens
Cheaper than 88% of models. Median price is $0.54/1M tokens.
Cost Calculator
Tokens per day1M
100K100M
Daily
$0.06
Monthly
$1.80
vs. Similar Models
Claude 4.5 Haiku (Non-reasoning)Q:-0.1
$2.00+3233%
Qwen: Qwen3 MaxQ:+0.2
$1.56+2500%
Claude 3.7 Sonnet (Non-reasoning)Q:-0.3
$6.00+9900%
MoonshotAI: Kimi K2 0905Q:-0.3
$1.07+1692%
Performance
307
tokens/sec
Faster than 96% of models
0.54
seconds
Faster than 84% of models
7.06
seconds
Faster than 44% of models
Market Median
94 tok/s
225% faster
Median TTFT
1.11s
52% faster
Throughput/Dollar
5112
tok/s per $/1M
Speed Comparison
Llama 3.1 Nemotron Instruct 70B
301 tok/s-2%
Nemotron 3 Nano Omni 30B A3B Reasoning
298 tok/s-3%
Nova Micro
289 tok/s-6%
Context Window
131K
tokens
Larger than 27% of models
Max Output
131K
tokens
100% of context
Benchmarks
MMLU-Pro
80.8%
GPQA Diamond
78.2%
HLE
18.5%
LiveCodeBench
87.8%
SciCode
38.9%
TerminalBench Hard
23.5%
MATH-500Not evaluated
AIMENot evaluated
AIME 2025
93.4%
IFBench
69.0%
Long Context Recall
50.7%
Tau2
65.8%
Market AverageTop Score
Open Source
apache-2.0120BGGUF / GPTQ / AWQ
Downloads
4.0M
Likes
4.9K
VRAM (FP16)
Multi-GPU
GPU
8x A100 / H100
