Private AI

Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Search
 
 
 
 
All conversations

Create

Chat
Studio

Explore

Models
Pricing
AI Apps
Blog

Library

Gallery
Workspace
Usage & billing
API
Suggestions and bugs
Use cases

|

Terms

|

Privacy

|

Refunds
Usage & billing
Teams
Referrals
Presets
Settings
FAQPress & mediaBrand assetsBenchmarksNano RPCEarn crypto
  1. Models
  2. Text
  3. OpenAI

GPT OSS 120B

An open-weight, 117B-parameter Mixture-of-Experts (MoE) language model designed for high-reasoning, agentic, and general-purpose production use cases. It activates 5.1B parameters per forward pass and is optimized to run on a single H100 GPU with native MXFP4 quantization. The model supports configurable reasoning depth, full chain-of-thought access, and native tool use, including function calling, browsing, and structured output generation.

  • Reasoning
  • Tool Calling
  • Structured Output

Added Feb 3, 2026

Model weights

Pricing

Auto routing · per 1M tokens
Input
$0.03
Output
$0.15
Cache read
$0.015
Compare provider prices

Specifications

Context window
128K
Max output
16.4K
Parameters
120B / 5.1B
Total / active

Benchmarks

Sourced from LMArena.

Arena Score

1351.7

Overall Rank

#220 / 421

Votes

30,133

Confidence Interval

1347.3 - 1356.1

Category Scores

Coding

#228 / 416

6,391 votes

1389.3

Math

#191 / 404

1,775 votes

1380.8

Longer Query

#268 / 399

6,546 votes

1324.3

Creative Writing

#271 / 419

3,892 votes

1278.8

Instruction Following

#238 / 421

7,842 votes

1323.5

Hard Prompts

#228 / 421

14,645 votes

1361.7

Additional Categories
22

Japanese

#161 / 288

574 votes

1323.9

Spanish

#189 / 303

850 votes

1368.7

Industry Mathematical

#191 / 401

1,534 votes

1383.5

German

#200 / 320

668 votes

1328.0

Polish

#200 / 244

1,734 votes

1331.1

French

#206 / 301

459 votes

1356.0

Korean

#212 / 289

680 votes

1266.7

Chinese

#216 / 395

2,273 votes

1376.4

Non English

#216 / 421

17,103 votes

1337.2

Industry Business And Management And Financial Operations

#217 / 414

5,352 votes

1352.2

Exclude Ties

#220 / 421

21,408 votes

1313.0

Expert

#220 / 371

1,393 votes

1357.2

Industry Medicine And Healthcare

#221 / 390

1,625 votes

1366.8

Russian

#222 / 385

2,136 votes

1336.4

Industry Software And It Services

#224 / 421

10,185 votes

1385.1

Industry Life And Physical And Social Science

#229 / 419

4,819 votes

1361.4

English

#235 / 421

12,991 votes

1361.4

Hard Prompts English

#240 / 419

6,461 votes

1365.9

Industry Legal And Government

#244 / 393

1,870 votes

1340.7

Multi Turn

#245 / 419

4,874 votes

1326.0

Industry Writing And Literature And Language

#248 / 420

6,627 votes

1309.8

Industry Entertainment And Sports And Media

#256 / 419

5,474 votes

1286.5

Published 2026-11-09 · Matched as gpt-oss-120b · CC BY 4.0

LMArena Dataset

Leaderboard results

BenchmarkScoreRank on NanoGPT
Epoch AI Benchmarking Hub

Epoch AI data, CC BY 4.0 · Updated Nov 13, 2026

Epoch Capabilities Index

140.2

#94 of 115

GPQA Diamond

75.8%

#88 of 119

OTIS Mock AIME 2024–2025

88.9%

#60 of 121

Mystery Game Puzzles

2.0%

#77 of 77

EQ-Bench

EQ-Bench data, MIT · Updated Nov 13, 2026

Creative Writing v3

956

#86 of 96

Longform Writing

35.8

#86 of 97

EQ-Bench 3

1073

#38 of 49

UGI Leaderboard

UGI Leaderboard data, Apache-2.0 · Updated Oct 16, 2026

UGI Score

19.6

#133 of 149

UGI Willingness

2.0

#120 of 149

UGI Writing

38.5

#96 of 149

Filtered and ranked to models on NanoGPT.

Providers

Choose or restrict providers with the API. Provider-selection prices include the selection markup. Provider selection guide · Pricing API (JSON, USD per 1K tokens)

Auto
Auto (ZDR)
$0.03$0.15$0.015N/AN/A
Auto
Auto (Anonymized)
$0.03$0.15$0.015N/AN/A
Runware
Runware
10% off
UNK
$0.0336$0.0302$0.147$0.132$0.0336$0.0302N/AN/A
Akash
Akash
FP4
GLB
$0.0315$0.196$0.0315N/AN/A
CoreWeave
CoreWeave
FP4
US
$0.0315$0.178$0.0315N/AN/A
DekaLLM
DekaLLM
BF16
ID
$0.0315$0.189$0.0315N/AN/A
Hyperfusion
Hyperfusion
MXFP4
AE
$0.0315$0.179N/AN/AN/A
io.net
io.net
FP4
US
$0.0315$0.157$0.0158N/AN/A
Venice
Venice
INT4
US
$0.0315$0.157$0.0315N/AN/A
PrimeIntellect
PrimeIntellect
FP4
UNK
$0.042$0.0892$0.0378N/AN/A
Crusoe
Crusoe
BF16
US
$0.0525$0.263$0.0525N/AN/A
Mancer
Mancer
FP8
$0.0578$0.263$0.0578N/AN/A
Sail Research (ASAP)
Sail Research (ASAP)
GLB
$0.063$0.42$0.0315N/AN/A
Sail Research (Standard)
Sail Research (Standard)
FP4
GLB
$0.063$0.42$0.0315N/AN/A
Google Vertex
Google Vertex
US
$0.0945$0.378N/AN/AN/A
BaseTen
BaseTen
FP4
US
$0.105$0.525$0.105N/AN/A
Parasail
Parasail
FP4
US
$0.105$0.787$0.0578N/AN/A
SambaNova
SambaNova
US
$0.147$0.998N/AN/AN/A
Uomi
Uomi
GLB
$0.149$0.662N/AN/AN/A
Aster
Aster
$0.157$0.63N/AN/AN/A
Aster Fast
Aster Fast
50% off
$0.315$0.157$1.26$0.63N/AN/AN/A
Groq
Groq
US
$0.157$0.63$0.0787N/AN/A
Nebius
Nebius
FP4
GLB
$0.157$0.63$0.105N/AN/A
Ollama
Ollama
UNK
$0.157$0.63$0.0147N/AN/A
SiliconFlow
SiliconFlow
FP8
SG
$0.157$0.63$0.0787N/AN/A
Tinfoil
Tinfoil
TEE verified
US
$0.157$0.63N/AN/AN/A
Together
Together
US
$0.157$0.63N/AN/AN/A
IONOS
IONOS
EU
$0.179$0.746N/AN/AN/A
SCX (Global)
SCX (Global)
FP8
GLB
$0.179$0.578N/AN/AN/A
Cerebras
Cerebras
FP16
US
$0.367$0.787N/AN/AN/A
Privacy:ZDR, prompts not retainedNo training on promptsRetention and training unclear
Batch API
ProviderInputOutputCached input
Auto
Autoopenai/gpt-oss-120b
$0.03/M$0.15/M$0.015/M

Batch rows for this model use normal routing and cost the same as standard API requests. Set provider in a row to choose a provider.

Related text models

Compare GPT OSS 120B with similar models from the same provider or model family.

GPT OSS Safeguard 20B

openai/gpt-oss-safeguard-20b

gpt-oss-safeguard is a first open weight reasoning model specifically trained for safety classification tasks to help classify text content based on customizable policies. As a fine-tuned version of gpt-oss, gpt-oss-safeguard is designed to follow explicit written policies that you provide. This enables bring-your-own-policy Trust & Safety AI, where your own taxonomy, definitions, and thresholds guide classification decisions. Well crafted policies unlock gpt-oss-safeguard's reasoning capabilities, enabling it to handle nuanced content, explain borderline decisions, and adapt to contextual factors.

GPT OSS 20B

openai/gpt-oss-20b

An open-weight 21B parameter model released under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for lower-latency inference and deployability on consumer or single-GPU hardware. The model is trained in OpenAI's Harmony response format and supports reasoning level configuration, fine-tuning, and agentic capabilities including function calling, tool use, and structured outputs.

GPT 6.1 Terra Pro

openai/gpt-6.1-terra-pro

GPT-6.1 Terra Pro uses the same underlying model as GPT-6.1 Terra with Pro reasoning mode enabled for higher-quality responses on complex tasks.

GPT 6.1 Terra

openai/gpt-6.1-terra

GPT-6.1 Terra is a fast, efficient model for everyday work, with strong visual reasoning, clearer writing, and capable coding at half GPT-6.1 Sol’s standard token prices. It supports text and image input, tool calling, structured outputs, and optional reasoning from none through max.

GPT 6.1 Sol

openai/gpt-6.1-sol

GPT-6.1 Sol delivers near-Astra performance at a lower cost for complex coding, computer use, and professional work. It supports text and image input, tool calling, structured outputs, and reasoning from low through max.

GPT 6.1 Sol Pro

openai/gpt-6.1-sol-pro

GPT-6.1 Sol Pro uses the same underlying model as GPT-6.1 Sol with Pro reasoning mode enabled for higher-quality responses on complex tasks.