GPT-5.4 nano

+ contribute

GPT-5.4 nano is a closed-weights language model from OpenAI, released March 2026.

68.6±0.8#99 of 719 LLMsfrom 23 benchmark results across 7 sources

$ how it compares

The LLMs ranked either side of GPT-5.4 nano, out of 828 with a score, then more from OpenAI.

#	model	score	where	RAM	released
107	GLM-4.7 Z.ai	69±1	api $1.00/M +local	~58GB	2025-12
108	GPT-5.5 Instant OpenAI	69±1	comparison only		2026-05
109	GPT-5.4 nano OpenAI	69±1	api $0.46/M		2026-03
110	Claude Sonnet 4.5 Anthropic	69±1	api $6.00/M		2025-09
111	Qwen3.5 27B Alibaba	69±0	api $0.84/M +local	~18GB	2026-02
	more from OpenAI
1	GPT-6 Astra OpenAI	87±1	api $20.00/M		2026-09
3	GPT-6.1 Sol OpenAI	85±1	api $4.00/M		2026-09
7	GPT-6 Sol OpenAI	83±1	api $4.00/M		2026-09

see all LLMs · find the ones that fit your machine

$ what it's good at

agent tasks▁▁▃▆█▆▅▅▅▄▃▁ #91 of 300 · 18% of its score · 7 tests

Beats Grok 4.20 (0309 reasoning), GPT-5.1, Grok 4.3 here, though they rank higher overall. Trails Hy3 preview (295B-A21B), Qwen3.6 35B-A3B, MiniMax-M2.7 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Terminal-Bench 2.0 lab	46.3% ✎	██···██··███ beats 1/6	very hard · avg 14%	▇▇ 4.3%	at its level
Toolathlon (Tool Decathlon) lab	35.5% ✎	▃··▃··▃··▃█▅ beats 1/8	very hard · avg 13%	▇ 3.7%	at its level
Humanity's Last Exam with tools lab	37.7% ✎	▃·▃·▅····▅·█ beats 1/8	very hard · avg 12%	▇ 3.3%	at its level
OSWorld-Verified lab	39.0% ✎	█▄·····▄·▄▄▄ beats 1/6	hard · avg 17%	▇ 2.4%	at its level
τ²-bench Telecom lab	92.5% ✎	▃·▅▅▃▃··█▃▃▅ beats 7/13	hard · avg 16%	▇ 2.3%	sharp · at its level · noisy
SWE-bench Pro lab	52.4% ✎	▄·····▄·▄▄▄█ beats 1/6	hard · avg 28%	▇ 1.5%	blunt · at its level
T1-Bench multi-domain conversational agent pass^3	7.2% ✎	▃▂··▂·▂·▂▂█▄ beats 0/11	hard · avg 25%	▇ 0.0%	blunt · at its level · noisy

coding▁▁▁▃▆██▇▆▄▃▁ #83 of 458 * · 6.2% of its score · 2 tests

Beats Grok 4.20 (0309 reasoning), GPT-5.4 mini, GPT-5.1 here, though they rank higher overall. Trails Llama 3.1 Nemotron Ultra 253B v1 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Luau Leetcode fully solved (Torpedo Software)	64.9% ✎	▁▁▁▁·▁▂▃▅▃█▇ beats 17/49	hard · avg 18%	▇▇ 4.7%	at its level
SWE-bench Pro lab	52.4% ✎	▄·····▄·▄▄▄█ beats 1/6	hard · avg 28%	▇ 1.5%	blunt · at its level

chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #109 of 393 * · 5.4% of its score · 1 test

Beats LongCat 2.0 (1.6T-A48B) here, though they rank higher overall. Trails Nemotron-Cascade-8B, Gemini 2.5 Flash, Mistral Large 3 (675B-A41B) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
LMArena (text)	1401.3 ✎	▁▁▁▂▃▃▃▆▆▇█▃ beats 246/379	avg 1268	▇▇ 5.4%	blunt

math▁▂▃▅▆▅▄█▄▁▁▁ #62 of 398 · 34% of its score · 3 tests

Beats Qwen3.6 Max Preview, Claude Sonnet 4.6, Claude Opus 4.5 here, though they rank higher overall. Trails INTELLECT-3, GPT-5 mini here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
FrontierMath tiers 1–3 v2	44.9% ✎	▂··▁▁▃▄▄█▅▃▃ beats 31/82	very hard · avg 1.2%	▇▇▇▇▇▇▇▇▇▇ 24%	sharp · at its level
FrontierMath tier 4 v2	12.2% ✎	▁▂▁▃█▅▂▂▁▁▁▁ beats 13/63	very hard · avg 0.1%	▇▇▇ 7.4%	sharp · too hard for it
OTIS mock AIME 2024–2025	87.8% ✎	▃▃▄▃▃▃▇▇█▃▂▅ beats 136/193	very hard · avg 8.5%	▇ 3.2%	sharp · noisy

vision & documents▁▁▃▂▁▃▄▂▆▇█▄ #80 of 147 * · 8.4% of its score · 2 tests

Beats DeepSeek V4 Flash Vision (Exp) here, though they rank higher overall. Trails GPT-4.1 mini, Gemini 2.5 Flash, GPT-4.1 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
MMMU Pro lab	66.1% ✎	▅·····▃·▃·█▃ beats 0/7	hard · avg 40%	▇▇ 4.6%	at its level
Arena vision (style controlled)	1199.3 ✎	▁▁▂▂▂▂▄▅▅▆██ beats 69/145	avg 1068	▇▇ 3.8%	blunt

web search▂▁▂▆▄█▃▃▁▁·▁ #69 of 80 * · 3.3% of its score · 1 test

Beats DeepSeek V4 Flash (Preview, Apr 2026), Nemotron 3 Ultra 550B-A55B here, though they rank higher overall. Trails GPT-4.1 nano, GPT-4o mini (2024-07-18), Mistral Small 3.1 24B (2503) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Humanity's Last Exam with tools lab	37.7% ✎	▃·▃·▅····▅·█ beats 1/8	very hard · avg 12%	▇ 3.3%	at its level

other tests9 not in a task · 29% of its score

test	result	vs the others in its run	difficulty	counts	why
GPQA Diamond	78.5% ✎	▁▁▁▁▂▄▆▄▅▇█▅ beats 122/208	hard · avg 40%	▇▇▇ 7.4%	sharp
Humanity's Last Exam lab	24.3% ✎	▃··▃·▃▃···▃█ beats 1/7	very hard · avg 4.4%	▇▇ 4.7%	typical
ARC-AGI-1 public evaluation set (ARC Prize model_baseline, pass@2)	51.6% ✎	▁·▇▃▁·█▇▁·▇▃ beats 9/27	very hard · avg 4.1%	▇▇ 4.4%	sharp · at its level · noisy
ARC-AGI-2 public evaluation set (ARC Prize model_baseline, pass@2)	3.5% ✎	█··▃▅▁▃▁▃·▁▃ beats 7/24	very hard · avg 0.3%	▇ 3.7%	sharp · noisy
Mystery Game Puzzles (Epoch)	5.0% ✎	█··▂▃▇▄█▂▃▁▁ beats 11/74	very hard · avg 9.4%	▇ 2.8%	sharp · too hard for it · noisy
Chess Puzzles (Epoch)	30.0% ✎	█▁▁▁▁▂▁▄▃▂▁▁ beats 103/135	very hard · avg 5.9%	▇ 2.6%	sharp · noisy
SimpleQA Verified	11.7% ✎	▁▁▃▂▃▃▆▆█▃▂▅ beats 4/78	very hard · avg 8.0%	▇ 1.9%	at its level
RobloxQA v2.0 multiple choice (Torpedo Software)	89.0% ✎	▁▁▃▁█▄▃▃▆▂▂▂ beats 10/49	easy · avg 78%	▇ 1.8%	typical

1 more test

test	result	vs the others in its run	difficulty	counts	why
Vectara hallucination leaderboard (HHEM-2.3, factual consistency rate)	96.9% ✎	▁▂▁▂▅█▂▃▄▃▁▃ beats 98/98	very easy · avg 90%	▇ 0.0%	blunt

$ where to run it

Through an API

api · $0.46/M · $0.2 in / $1.25 out per 1M tokens · 400k context · platform.openai.com ↗ · price sourceStandard base token rate from models.dev. Provider-specific tiers, long-context surcharges, caching and promotions may differ.

$ how the score was worked out · Epoch AI — internal evaluatio… 49%, OpenAI vendor-reported benchm… 27%, LMArena leaderboard dataset 9% and 4 more

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

source	results	share of the score	licence
Epoch AI — internal evaluations	7	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 49%	CC BY 4.0
OpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards) · lab-reported	8	▇▇▇▇▇▇▇▇ 27%	Factual vendor-reported score (not licensed content)
LMArena leaderboard dataset	2	▇▇▇ 9%	CC BY 4.0
ARC Prize Foundation model_baseline outputs on ARC-AGI-1/2 public evaluation sets	2	▇▇ 8%	MIT
Roblox LLM Leaderboard (Torpedo Software): RobloxQA v2.0 and Luau Leetcode	2	▇▇ 6%	MIT
T1-Bench: Benchmarking Multi-Scenario Agents in Real-World Domains (arXiv:2606.11070v1)	1	▇ 0%	CC BY 4.0
Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)	1	▇ 0%	Apache-2.0
prior · pull toward the middle		▇ 0%
weighted average		= 68.6