GPT-5

+ contribute

GPT-5 is a closed-weights language model from OpenAI, released August 2025.

72.4±0.6#60 of 719 LLMsfrom 53 benchmark results across 28 sources

$ how it compares

The LLMs ranked either side of GPT-5, out of 828 with a score, then more from OpenAI.

#	model	score	where	RAM	released
63	LongCat 2.0 (1.6T-A48B) Meituan	73±2	api $1.30/M +local	~2155GB	2026-06
64	Gemini 3 Pro (Preview) Google	73±1	comparison only		2025-11
65	GPT-5 OpenAI	72±1	api $3.44/M		2025-08
66	Qwen3.6 Max Preview Alibaba	72±1	api $2.92/M		2026-04
67	Claude Sonnet 4.6 Anthropic	72±1	api $6.00/M		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▁▁▃▆█▆▅▅▅▄▃▁ #81 of 300 · 20% of its score · 10 tests

Trails MiniMax-M2.7, Step 3.5 Flash (196B-A11B), MiMo-V2.5 (310B-A15B) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
SWE-bench Verified	73.6% ✎	▁··▁·▁▂▁▄█▄▁ beats 13/29	hard · avg 19%	▇▇ 5.6%	sharp
Terminal-Bench 2.0 with Terminus 2 agent (benchmark authors' runs)	35.2% ✎	▃··▃·▃▃██▅▃▅ beats 10/14	very hard · avg 5.8%	▇ 3.1%	sharp · at its level
τ²-bench Telecom lab +1	96.7% ✎	▃·▅▅▃▃··█▃▃▅ beats 11/13	very hard · avg 13%	▇ 2.5%	sharp · several runs
BrowserGym WorkArena-L2 (AgentLab GenericAgent)	69.4% ✎	▃▅▃·▅▃▃··█·▃ beats 11/11	very hard · avg 3.9%	▇ 2.5%	sharp · at its level · noisy
ContractBench success rate (33 observation-contract tool tasks)	70.7% ✎	▆··▆▄▂▄▂▆▆█▄ beats 21/23	very hard · avg 14%	▇ 2.4%	sharp · at its level
τ²-bench Retail	81.6% ✎	▁····▂▁▃▆█▃▃ beats 43/44	medium · avg 41%	▇ 1.3%	typical
τ²-bench Airline	62.5% ✎	▁▁·▂▁▃▂▁█▅▂▃ beats 35/42	hard · avg 29%	▇ 1.1%	at its level
BrowserGym MiniWoB (AgentLab GenericAgent)	71.5% ✎	▅····▅▅▃▃·▃█ beats 11/11	medium · avg 55%	▇ 0.8%	blunt

2 more tests

test	result	vs the others in its run	difficulty	counts	why
BFCL V4 (2025-11-03 archive, models absent from the December archive)	61.6% ✎	▃··▅▃▃·▅·▃·█ beats 7/10	hard · avg 34%	▇ 0.6%	at its level
τ²-Bench-Verified (GPT-5.1 user)	79.9% ✎	▂··▄···▂▄▄█▄ beats 10/13	medium · avg 45%	▇ 0.5%	typical

coding▁▁▁▃▆██▇▆▄▃▁ #75 of 458 · 20% of its score · 6 tests

Beats GPT-5.2 here, though they rank higher overall. Trails GLM-4.7, Gemini 3.5 Flash-Lite, MiMo-V2.5 (310B-A15B) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
SWE-bench Verified	73.6% ✎	▁··▁·▁▂▁▄█▄▁ beats 13/29	hard · avg 19%	▇▇ 5.6%	sharp
Arena WebDev (Code Arena)	1417.1 ✎	▁▂▄▃▅█▄▇▆▄▁▁ beats 50/119	avg 981	▇▇ 5.2%	sharp
Aider polyglot (225 exercises, two attempts)	88.0% ✎	▂·▂▃▅▂▃█▆▄▂▃ beats 49/49	very hard · avg 8.5%	▇▇ 4.1%	sharp
Luau Leetcode fully solved (Torpedo Software)	78.8% ✎	▁▁▁▁·▁▂▃▅▃█▇ beats 31/49	hard · avg 18%	▇ 2.3%	at its level
VisEditBench strict pass rate (1,395 visualization-code editing tasks)	65.7% ✎	▄··█·▄··▄█·▄ beats 6/7	hard · avg 34%	▇ 1.4%	at its level
WebApp1K-Duo-React pass@1 (onekq-ai)	79.4% ✎	▁··▁·▁▁▁▆█▇█ beats 45/45	hard · avg 32%	▇ 1.0%	typical

chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #72 of 393 * · 2.9% of its score · 1 test

Beats LongCat 2.0 (1.6T-A48B) here, though they rank higher overall. Trails Gemma 4 26B-A4B, Nemotron Cascade 2 30B-A3B, Gemma 4 31B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
LMArena (text)	1434.5 ✎	▁▁▁▂▃▃▃▆▆▇█▃ beats 292/379	avg 1268	▇ 2.9%	blunt

math▁▂▃▅▆▅▄█▄▁▁▁ #45 of 398 · 26% of its score · 6 tests

Beats Grok 4.7, GLM-5.3-Flash, GPT-5 Pro here, though they rank higher overall. Trails A.X K2 (688B-A33B) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
FrontierMath tiers 1–3 v2	55.4% ✎	▂··▁▁▃▄▄█▅▃▃ beats 39/82	very hard · avg 1.2%	▇▇▇▇▇▇ 15%	sharp · at its level
FrontierMath tier 4 v2	22.0% ✎	▁▂▁▂█▅▂▂▁▁▁▁ beats 22/63	very hard · avg 0.1%	▇▇▇ 8.3%	sharp
MathNet-Solve-Test micro-average accuracy (6,400 olympiad problems)	69.3% ✎	▃▃···█▃·▃·▃▃ beats 7/8	hard · avg 15%	▇ 1.0%	at its level
OTIS mock AIME 2024–2025	91.4% ✎	▃▃▄▃▃▃▇▇█▃▂▅ beats 150/193	very hard · avg 8.5%	▇ 0.9%	sharp · noisy
MATH level 5	98.1% ✎	▂▂▅▆▄█▆▃▅▂▃▄ beats 92/92	hard · avg 32%	▇ 0.7%	sharp · maxed out
FormalProofBench (200 graduate-level Lean 4 problems, agentic)	17.0% ✎	▃▅·▃··▅▃█··▃ beats 7/10	very hard · avg 3.0%	▇ 0.6%	noisy

vision & documents▁▁▃▂▁▃▄▂▆▇█▄ #68 of 147 · 7.7% of its score · 8 tests

Beats DeepSeek V4 Flash Vision (Exp) here, though they rank higher overall. Trails Gemma 4 26B-A4B, Claude Haiku 4.5, Qwen3.6 Flash here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Arena vision (style controlled)	1210.2 ✎	▁▁▂▂▂▂▄▅▅▆██ beats 77/145	avg 1068	▇ 2.0%	blunt
ChartArena chart parsing (English, mAP_high average)	45.6% ✎	▁·▁·▃▃▁▇▃█·▄ beats 16/22	hard · avg 27%	▇ 1.6%	blunt · at its level
VisEditBench strict pass rate (1,395 visualization-code editing tasks)	65.7% ✎	▄··█·▄··▄█·▄ beats 6/7	hard · avg 34%	▇ 1.4%	at its level
CharXiv descriptive questions +1	91.3% ✎	▄····▄···█▄▄ beats 4/5	easy · avg 70%	▇ 1.2%	maxed out · several runs
CharXiv (reasoning questions)	71.5% ✎	▄·····▄▄·▄·█ beats 5/5	hard · avg 33%	▇ 0.8%	typical
Molmo2-VideoCount accuracy (val)	35.8% ✎	▃··▃·▃····█▃ beats 5/6	hard · avg 22%	▇ 0.3%	blunt · at its level
Ai2 Molmo2 human preference Elo (video/image QA, 2026)	1031.0 ✎	▄··▄··▄··▄█▄ beats 1/6	avg 975	▇ 0.2%	blunt
Molmo2-CapTest dense video captioning F1	50.1% ✎	█···█·████·█ beats 5/6	hard · avg 32%	▇ 0.1%	blunt · at its level

web search▂▁▂▆▄█▃▃▁▁·▁ #41 of 80 * · 1.2% of its score · 1 test

Beats Gemini 3.1 Pro (Preview), Qwen3.7 Max, Gemini 3.5 Flash here, though they rank higher overall. Trails GPT-4o mini (2024-07-18), Mistral Small 3.1 24B (2503), Llama 4 Scout (109B-A17B) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Arena search systems (style controlled)	1182.9 ✎	▁·▁▁▂▂▃█▅▂▁▁ beats 10/30	avg 1055	▇ 1.2%	blunt

other tests20 not in a task · 29% of its score

test	result	vs the others in its run	difficulty	counts	why
GPQA Diamond +1	86.2% ✎	▁▁▁▁▂▄▆▄▅▇█▅ beats 152/208	hard · avg 39%	▇▇▇ 6.9%	sharp · several runs
EBR-bench (Epoch)	12.7% ✎	▄▂·▆█▂▆▄▆▆·▄ beats 6/23	very hard · avg 0.3%	▇▇ 4.1%	sharp
Mystery Game Puzzles (Epoch)	23.0% ✎	█··▂▃▇▄█▂▃▁▁ beats 40/74	very hard · avg 9.4%	▇ 2.7%	sharp · noisy
KCL-Essay vanilla	57.5% ✎	▁···▁▃▂█▂▃▃▃ beats 22/23	hard · avg 18%	▇ 2.2%	at its level
Chess Puzzles (Epoch)	37.0% ✎	█▁▁▁▁▂▁▄▃▂▁▁ beats 113/135	very hard · avg 5.9%	▇ 2.0%	sharp · noisy
RPC-Bench overall, text input (Z.ai)	37.5% ✎	▄··▄▄·▄▄█·▄▄ beats 8/8	very hard · avg 11%	▇ 1.4%	at its level
KCL-MCQA vanilla	53.9% ✎	▂···█▂▆▆██▆▂ beats 20/23	very hard · avg 30%	▇ 1.3%	at its level
SimpleQA Verified	50.1% ✎	▁▁▃▂▃▃▆▆█▃▂▅ beats 54/78	very hard · avg 8.0%	▇ 1.1%	at its level

12 more tests

test	result	vs the others in its run	difficulty	counts	why
LongSocialBench (full-context, 4-option MCQ over discussion threads)	52.9% ✎	▆▂·▄█·▂▂·▄▄▂ beats 15/16	very hard · avg 36%	▇ 1.1%	blunt · at its level
AraGen-v3 3C3H score (Inception, Claude Sonnet 4 judge)	84.3% ✎	▁▁·▁▂▂▄█▄▅▄▄ beats 65/66	hard · avg 31%	▇ 1.1%	typical
MMMLU lab	89.4% ✎	▂▁▁▂▃▂▄▄▅██▃ beats 44/46	medium · avg 64%	▇ 0.9%	small run
SimpleQA lab	55.0% ✎	▄▄▄▆█▆███·▂█ beats 32/32	very hard · avg 9.1%	▇ 0.8%	at its level · noisy
RobloxQA v2.0 multiple choice (Torpedo Software)	94.4% ✎	▁▁▃▁█▄▃▃▆▂▂▂ beats 27/49	easy · avg 78%	▇ 0.8%	typical
SOOHAK Challenge (340 problems, Avg@3)	26.4% ✎	█·█████··█·█ beats 7/7	very hard · avg 4.1%	▇ 0.6%	typical
DiagnosticIQ industrial maintenance MCQ (macro accuracy, 4 options)	65.9% ✎	▁··▃▁▁▁▅█▅▁▅ beats 19/24	medium · avg 52%	▇ 0.6%	blunt · at its level
HealthBench Hard lab	46.2% ✎	▁····▁▁▂█·▁▃ beats 13/13	hard · avg 15%	▇ 0.4%	blunt · at its level · small run
SOOHAK-Mini (702 problems, Avg@3)	72.2% ✎	▄··▄·▄·▄▄█·▄ beats 7/7	medium · avg 36%	▇ 0.4%	blunt · at its level
Open PT-LLM Leaderboard average (9 Portuguese tasks)	83.4% ✎	▃▁▁▁▁▁▁▁▂█▅▃ beats 271/276	medium · avg 73%	▇ 0.2%	noisy
Vectara hallucination leaderboard, 2025 dataset (HHEM-2.3, factual consistency rate)	98.6% ✎	▁▁▂▃▄▄█▆▃▃▂▁ beats 131/140	very easy · avg 94%	▇ 0.1%	blunt · maxed out · noisy
Vectara hallucination leaderboard (HHEM-2.3, factual consistency rate)	84.9% ✎	▁▂▁▃▅█▂▃▄▃▁▁ beats 10/98	very easy · avg 90%	▇ 0.0%	blunt

$ where to run it

Through an API

api · $3.44/M · $1.25 in / $10 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… 43%, LMArena leaderboard dataset 11%, OpenAI vendor-reported benchm… 5% and 25 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	10	▇▇▇▇▇▇▇▇▇▇▇▇▇ 43%	CC BY 4.0
LMArena leaderboard dataset	4	▇▇▇ 11%	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	2	▇ 5%	Factual vendor-reported score (not licensed content)
Aider polyglot benchmark runs	1	▇ 4%	Apache-2.0
τ²/τ³-bench leaderboard: Sierra's own evaluation runs	3	▇ 4%	MIT
Korean Canonical Legal Benchmark: Toward Knowledge-Independent Evaluation of LLMs’ Legal Reasoning Capabilities	2	▇ 3%	CC BY 4.0
ServiceNow BrowserGym leaderboard (AgentLab GenericAgent runs)	2	▇ 3%	MIT
Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces — benchmark authors' own agent-model trials	1	▇ 3%	CC BY 4.0
Roblox LLM Leaderboard (Torpedo Software): RobloxQA v2.0 and Luau Leetcode	2	▇ 3%	MIT
ContractBench: Can LLM Agents Preserve Observation Contracts? (arXiv:2605.17281v2)	1	▇ 2%	CC BY 4.0
OpenAI — GPT-5 System Card (arXiv:2601.03267v2) · lab-reported	3	▇ 2%	CC BY 4.0
ChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)	1	▇ 2%	CC BY 4.0
VisEditBench: Can Vision-Language Models Edit Visualization Code from Multimodal Feedback? (arXiv:2608.10408v1)	1	▇ 1%	CC BY 4.0
RPC-Bench leaderboard (Z.ai research paper comprehension)	1	▇ 1%	Apache-2.0
ServiceNow-AI/Apriel-1.6-15b-Thinker model card	2	▇ 1%	MIT
LongSocialBench: Do Long-Context LLMs Understand Online Discussion Threads? (arXiv:2610.04118v1)	1	▇ 1%	CC BY 4.0
Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)	1	▇ 1%	Apache-2.0
MathNet: a Global Multimodal Benchmark for Mathematical Reasoning and Retrieval — authors' own model runs	1	▇ 1%	CC BY 4.0
WebApp1K models leaderboard (onekq-ai)	1	▇ 1%	MIT
Soohak: A Mathematician-Curated Benchmark for Evaluating Research-level Math Capabilities of LLMs — authors' own model runs	2	▇ 1%	CC BY-SA 4.0
Baidu - ERNIE 5.0 Technical Report (arXiv:2602.04705v1)	1	▇ 1%	CC BY 4.0
Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding (Ai2, arXiv 2601.10611v4)	3	▇ 1%	CC BY 4.0
FormalProofBench (Vals AI, arXiv:2603.26996v1)	1	▇ 1%	CC-BY-4.0
Berkeley Function Calling Leaderboard result archive	1	▇ 1%	Apache-2.0
DiagnosticIQ: A Benchmark for LLM-Based Industrial Maintenance Action Recommendation from Symbolic Rules (arXiv:2605.08614v1)	1	▇ 1%	CC BY 4.0
Amazon AGI — τ²-Bench-Verified	1	▇ 0%	MIT
Open Portuguese LLM Leaderboard (archived)	1	▇ 0%	Apache-2.0
Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)	2	▇ 0%	Apache-2.0
prior · pull toward the middle		▇ 0%
weighted average		= 72.4