Automated page. Written and translated automatically by AI and checked against the sources by software; no human editor. SI-IQ is an index built from public benchmark results, not an IQ test. How it works

How smart is gpt-oss-120b?

106 SI-IQ · 90% interval 99–112
Rank 64 of 80 ranked models · Plausible rank range 47–71

gpt-oss-120b is a model from OpenAI, released on Aug 5, 2025. Its SI-IQ is 106 (90% interval 99–112), estimated from 5 benchmark results by 3 independent evaluators.

Facts

Developer: OpenAI
Released: Aug 5, 2025
Access: Open weights (restricted licence)
Benchmarks used: 5 results from 3 evaluators

Ability by area (SI-IQ, 90% interval)

Every benchmark result we used

Each score is one value exactly as the source published it: for Epoch AI the best of its run settings, for OpenRouter the median hosting provider (the developer's own hosting when it ran the test). “Implied SI-IQ” is the ability that this single result points to; the overall SI-IQ is their precision-weighted average.

Benchmark	Area	Score	Implied SI-IQ	Evaluated by	As listed by the source	Retrieved	Source
Mystery Game Puzzles	Reasoning	2.0%	70	Epoch AI	`gpt-oss-120b_medium`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
OTIS Mock AIME 2024–2025	Math	88.9%	115	Epoch AI	`openai/gpt-oss-120b_high`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
Aider Polyglot	Coding	41.8%	91	Aider	`gpt-oss-120b (high)`	Oct 18, 2026	Aider LLM Leaderboards · data file
GPQA Diamond	Knowledge	75.8%	102	Epoch AI	`openai/gpt-oss-120b_high`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
GPQA Diamond (OpenRouter run)	Knowledge	75.8%	105	OpenRouter	`openai/gpt-oss-120b` Together · median over providers (22 of 22 rows), lower median	Oct 18, 2026	OpenRouter benchmark datasets · data file
τ²-bench Airline (OpenRouter run)	Agents	63.3%	109	OpenRouter	`openai/gpt-oss-120b` Venice · median over providers (18 of 18 rows), lower median	Oct 18, 2026	OpenRouter benchmark datasets · data file

Human preference (Arena)

On the Arena leaderboard, where people vote between two anonymous answers, gpt-oss-120b has a rating of 1,352 (1,347–1,356) from 30,206 votes (published Oct 15, 2026). Human preference is a different measure from benchmark ability and is not part of SI-IQ. Arena (Hugging Face)

Models with similar SI-IQ

Model	SI-IQ
GPT-5 nanoOpenAI	105	gpt-oss-120b vs GPT-5 nano
o3-miniOpenAI	105	gpt-oss-120b vs o3-mini
Claude Haiku 4.5Anthropic	104	gpt-oss-120b vs Claude Haiku 4.5
MiniMax-M3MiniMax	108	gpt-oss-120b vs MiniMax-M3
Gemini 2.5 Pro (Jun 2025)Google DeepMind	108	gpt-oss-120b vs Gemini 2.5 Pro (Jun 2025)

Computed with SI-IQ method version 1.0.0. Read the full methodology

Data sources: Aider polyglot leaderboard data (Apache-2.0); Arena leaderboard dataset on Hugging Face (CC BY 4.0); Epoch AI, ‘Capabilities & benchmarking’ (CC BY 4.0); Source: OpenRouter (openrouter.ai), licensed under CC BY 4.0: reuse and republish with attribution to OpenRouter..