GPT-4 (0613)

+ contribute

GPT-4 (0613) is a closed-weights language model from OpenAI, released June 2023. Licence: proprietary.

Catalogue identity notes

Comparison model from the Epoch AI internal-evaluation CSVs (Epoch release date 2023-06-13). Dated API snapshot gpt-4-0613 (default snapshot of 'gpt-4'). OpenAI's deprecations page schedules its shutdown for October 23, 2026 (announced 2026-04-22).

51.3±0.6#322 of 719 LLMsfrom 27 benchmark results across 15 sources

$ how it compares

The LLMs ranked either side of GPT-4 (0613), out of 828 with a score, then more from OpenAI.

#	model	score	where	RAM	released
388	Cogito v1 Preview Llama 8B Deep Cogito	51±5	local	~6.2GB	2025-03
389	Hermes 3 Llama 3.1 405B Nous Research	51±1	local	~114GB	2024-08
390	GPT-4 (0613) OpenAI	51±1	api $37.50/M		2023-06
391	OpenCoder-8B-Instruct INF (infly)	51±1	local	~6.4GB	2024-11
392	Command A Translate (08-2025) Cohere	51±5	api price unknown		2025-08
	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

coding▁▁▁▃▆██▇▆▄▃▁ #221 of 458 · 24% of its score · 6 tests

Beats Qwen3.5 4B, Qwen3-VL 32B Instruct, Solar Open 100B (102B-A12B) here, though they rank higher overall. Trails QwQ-32B-Preview, Ministral 3 3B Reasoning 2512 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
HumanEval	82.3% ✎	▁█▄▅▇█·▄▇▅▄▄ beats 41/42	easy · avg 78%	▇▇▇▇ 9.2%	sharp
LiveCodeBench official (Jan 1–Nov 1, 2024)	32.5% ✎	▁▁·▁▂▄▆█▅▂·▂ beats 17/31	hard · avg 26%	▇▇▇ 6.4%	sharp · at its level
Aider code editing (133 Exercism Python exercises, two attempts)	67.7% ✎	▁·▁▂▄▁▆▄█▆▂▂ beats 44/62	medium · avg 54%	▇▇ 5.3%	at its level
CRUXEval-I	69.8% ✎	▁··▁▅▂█▇▁··▃ beats 22/22	medium · avg 54%	▇ 1.1%	at its level
CRUXEval-O	68.7% ✎	▁·▂▃▁█▁▂···▃ beats 22/22	medium · avg 53%	▇ 1.1%	blunt · at its level
Tulu 2: CodexEval pass@10	87.0% ✎	▃▃▃·▆▃▃█▃▂▂▅ beats 23/25	very easy · avg 86%	▇ 1.0%	sharp

chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #262 of 393 · 28% of its score · 6 tests

Beats Ministral 3 14B Reasoning 2512, Ministral 3 14B, Phi-4 14B here, though they rank higher overall. Trails Flan-T5 XXL, Granite 4.0 H 1B, Granite 4.0 1B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
AlpacaEval 2 length-controlled	30.2% ✎	▁▂▂▃▆▆█▆▄▄▁▁ beats 78/122	medium · avg 39%	▇▇▇▇▇ 12%	at its level
Arena-Hard (v0.1, GPT-4-1106 judge)	37.9% ✎	▂▁▁▅█▅▄▃▃▅▃▁ beats 33/60	medium · avg 52%	▇▇▇ 7.6%	sharp · at its level
Polyrating rating (bias-aware re-fit of Chatbot Arena votes, mid-2024)	1191.0 ✎	▂▁▂▁▃▇█▆▇▄▂▂ beats 80/95	avg 1201	▇ 2.8%	blunt
LMArena (text)	1276.2 ✎	▁▁▁▂▃▃▃▆▆▇█▃ beats 107/379	avg 1268	▇ 2.8%	blunt
MT-Bench (GPT-4 single-answer grading, LMSYS runs)	9.1800 MT-Bench score (1-10) ✎	▆▂▂▆▆▄██▂▄▄▆ beats 27/28	avg 8.819 MT-Bench score (1-10)	▇ 2.5%	sharp · noisy
AlpacaEval 1.0 win rate (vs text-davinci-003, GPT-4 judge)	93.8% ✎	▂▂·▆·▇▃▆█▆▄▂ beats 38/45	very easy · avg 96%	▇ 0.2%	sharp · maxed out · noisy

math▁▂▃▅▆▅▄█▄▁▁▁ #308 of 398 · 8.9% of its score · 3 tests

Trails Gemma 7B (base), DeepSeek R1 Distill Qwen 1.5B, Phi-3.5-vision-instruct here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
MATH level 5	23.0% ✎	▂▂▅▆▄█▆▃▅▂▃▄ beats 21/92	hard · avg 32%	▇▇▇ 6.7%	sharp · at its level
OTIS mock AIME 2024–2025	1.1% ✎	▃▃▄▃▃▃▇▇█▃▂▅ beats 8/193	very hard · avg 8.5%	▇ 1.1%	sharp · noisy
GSM8K	95.0% ✎	▆▂▅█▅▅▃▆···▃ beats 25/25	easy · avg 73%	▇ 1.1%	sharp

other tests12 not in a task · 39% of its score

test	result	vs the others in its run	difficulty	counts	why
LiveBench 2024-07-26	45.6% ✎	▁▁▁▂▂▃▂█▆▃▅▃ beats 48/61	medium · avg 41%	▇▇▇▇▇ 12%	at its level
EvoEval (overall)	66.2% ✎	▂▂▃▄▄█▁▄▅▃▁▃ beats 42/42	medium · avg 55%	▇▇▇ 8.6%	at its level
BIG-Bench Hard	89.1% ✎	▁·····▂█▂▅·▃ beats 25/25	easy · avg 77%	▇▇ 3.8%	sharp · at its level
GPQA Diamond	30.7% ✎	▁▁▁▁▃▄▆▄▅▇█▅ beats 18/208	hard · avg 40%	▇ 3.6%	sharp
BigCodeBench (Instruct)	45.9% ✎	▃▂·▆▆██▆▅▃▃▃ beats 31/33	medium · avg 41%	▇ 3.6%	at its level
BigCodeBench (Complete)	48.4% ✎	▂▁▁▄▂▅▇█▆▂▂▂ beats 47/58	medium · avg 50%	▇ 3.4%	at its level
MMLU	81.4% ✎	▁▁▃▃▄█▄▁█··▃ beats 25/25	easy · avg 74%	▇ 1.5%	at its level
EQ-Bench v1 (60 questions, paper run)	62.5% ✎	▂··▄·▂▆▄▂█·▃ beats 14/14	medium · avg 57%	▇ 1.2%	at its level

4 more tests

test	result	vs the others in its run	difficulty	counts	why
Tulu 2: TyDiQA gold passage, 1-shot F1	65.2% ✎	▂··▁▃▂▂▂█▅▁▃ beats 25/25	medium · avg 61%	▇ 0.5%	blunt · at its level
Vectara hallucination leaderboard, 2025 dataset (HHEM-2.3, factual consistency rate)	98.2% ✎	▁▁▂▃▄▄█▆▃▃▂▁ beats 125/140	very easy · avg 94%	▇ 0.2%	blunt · maxed out · noisy
Chess Puzzles (Epoch)	4.0% ✎	█▁▁▁▁▂▁▄▂▂▁▁ beats 13/135	very hard · avg 5.9%	▇ 0.1%	sharp · too hard for it · noisy
Mystery Game Puzzles (Epoch)	12.0% ✎	█··▃▃▇▄█▂▃▁▁ beats 17/74	very hard · avg 9.4%	▇ 0.1%	sharp · too hard for it · noisy

$ where to run it

Through an API

api · $37.50/M · $30 in / $60 out per 1M tokens · 8k context · developers.openai.com ↗ · price sourceUSD per 1M tokens, OpenAI API standard processing, model page 'gpt-4' whose default snapshot is gpt-4-0613: input $30, output $60; cached-input, Batch and other tiers excluded. Currency: USD. OpenAI official model page, observed 2026-09-28.

$ how the score was worked out · Top Leaderboard Ranking = Top… 18%, AlpacaEval 1.0 and 2.0 leader… 12%, LiveBench model judgments (HF… 12% and 12 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
Top Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runs	2	▇▇▇▇▇ 18%	CC BY 4.0
AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)	2	▇▇▇▇ 12%	Apache-2.0
LiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasets	1	▇▇▇▇ 12%	Apache-2.0 (stated in the LiveBench datasheet)
Epoch AI — internal evaluations	5	▇▇▇ 12%	CC BY 4.0
Tulu 2 — base and instruction-tuned evaluations	5	▇▇ 8%	CC BY 4.0
Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)	1	▇▇ 8%	Apache-2.0
BigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4	2	▇▇ 7%	CC BY 4.0
LiveCodeBench official website results (2024-12-26)	1	▇▇ 6%	CC-BY-SA-4.0
Aider code editing benchmark runs (old edit leaderboard)	1	▇▇ 5%	Apache-2.0
Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)	1	▇ 3%	CC-BY-4.0
LMArena leaderboard dataset	1	▇ 3%	CC BY 4.0
LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table	1	▇ 2%	Apache-2.0
CRUXEval published evaluation results (authors' generations and scores)	2	▇ 2%	MIT
EQ-Bench: An Emotional Intelligence Benchmark for Large Language Models — authors' own runs	1	▇ 1%	CC BY 4.0
Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)	1	▇ 0%	Apache-2.0
prior · pull toward the middle		▇ 0%
weighted average		= 51.3