GPT-4 (Mar 2023): benchmark scores, price and ranking

GPT-4 (Mar 2023) from OpenAI ranks #664 of 743 LLMs on the LLMs Tiger aggregate benchmark score (43.7), based on 4 benchmark results captured up to 2026-10-13.

Organisation: OpenAI
Type: Frontier model
Open weights: No
Released: 2023-03-14 (source: epoch)
Cheapest API price: $30.00 in / $60.00 out per million tokens (via azure, observed 2026-10-13T20:47Z)
Aggregate rank: #664 of 743 (score 43.7)

Relative to other models, GPT-4 (Mar 2023) is strongest on Epoch ECI (#111 of 204) and weakest on OTIS Mock AIME (#185 of 189).

Across the 4 benchmarks where at least ten models are scored, GPT-4 (Mar 2023) is in the top 10% on 0 and in the top half on 0.

Its cheapest listed input price, $30.00 per million tokens, is lower than that of 1% of the 427 models with a listed price.

GPT-4 (Mar 2023) benchmark results

GPT-4 (Mar 2023) benchmark scores, ranked against all models measured on each benchmark
Benchmark	Score	Rank	Source	Captured
Epoch ECI	126	#111 of 204	Epoch AI Benchmarking Hub	2026-10-13
METR time horizon	5 min	#24 of 31	Epoch AI Benchmarking Hub	2026-10-13
GPQA Diamond	35.7%	#233 of 275	Epoch AI Benchmarking Hub	2026-10-13
OTIS Mock AIME	0.6%	#185 of 189	Epoch AI Benchmarking Hub	2026-10-13

About GPT-4 (Mar 2023)

GPT-4 (Mar 2023) is a frontier model from OpenAI. It was released on 2023-03-14, according to Epoch AI. The cheapest API listing we track charges $30.00 per million input tokens and $60.00 per million output tokens, through azure. We hold 4 benchmark results for it across 2 areas, and none is in the top tenth of its board.

Where GPT-4 (Mar 2023) stands

On our aggregate score, GPT-4 (Mar 2023) is #664 of 743 ranked models, in the lower half. Its score is 43.7; the scale is centred on 50, so it sits 6.3 points below the typical model. That beats 11% of the ranked field. Among the 88 ranked OpenAI models it is #79. The next one up is GPT-5 nano (high) at 43.8. The next one down is GPT-4.1 mini at 43.0.

GPT-4 (Mar 2023), benchmark by benchmark

GPT-4 (Mar 2023) lands below the middle on Epoch ECI: 126, #111 of 204, with 54% of the field scoring higher. Qwen 3.8 Max leads at 156; the closest result above is Llama 3.1-70B at 126. It covers Epoch Capabilities Index: Epoch's own aggregate of ~40 benchmarks. The board median is 127, so GPT-4 (Mar 2023) is 1 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

GPT-4 (Mar 2023) lands below the middle on METR time horizon: 5 min, #24 of 31, with 74% of the field scoring higher. Claude Mythos Preview (Early) leads at 1,045 min; the closest result above is GPT-4 Turbo Preview (January 2024) at 5 min. It covers length of task (minutes) done at 50% success. The board median is 23 min, so GPT-4 (Mar 2023) is 18 minutes below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

GPQA Diamond is a weak spot. GPT-4 (Mar 2023) scores 35.7%, #233 of 275, behind 84% of the models measured. The leader, GPT-6 Astra (max), reaches 95.8%. It covers expert-level science questions. The board median is 71.7%, so GPT-4 (Mar 2023) is 36.0 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

OTIS Mock AIME is a weak spot. GPT-4 (Mar 2023) scores 0.6%, #185 of 189, behind 97% of the models measured. The leader, Claude Fable 5.1 (max), reaches 100.0%. It covers competition maths (AIME 2024-25 style). The board median is 57.8%, so GPT-4 (Mar 2023) is 57.2 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Strengths and weaknesses

Reasoning, maths and science: weak, ahead of 9% of the field on average across 2 benchmarks. Its best board here is GPQA Diamond (#233 of 275) and its weakest is OTIS Mock AIME (#185 of 189). General knowledge and chat: uneven, ahead of 34% of the field on average across 2 benchmarks. Its best board here is Epoch ECI (#111 of 204) and its weakest is METR time horizon (#24 of 31).

Price and value

The median input price across the 427 priced models is $0.500 per million tokens, so GPT-4 (Mar 2023) costs more than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), Qwen2.5-Coder-3B-Instruct ($0.010, 46.5).

Release history

The OpenAI model released before it was GPT-3.5 on 2022-11-30, which scores 43.0. The next one was GPT-4 (0613) on 2023-06-13, at 50.8.

Reading these numbers

The 4 scores on this page come from 1 source: Epoch AI Benchmarking Hub (4). No official leaderboard supplied any of them, so treat them as third-party measurements. With 4 results the aggregate rank can move a lot when one new score arrives.

Models ranked nearby

#661 GPT-5 nano (high) (43.8)
#662 gemma-2-27b-it (43.8)
#663 Magistral Small 1.2 (43.8)
#665 gemma-2-27b (43.7)
#666 mistral-medium-2508 (43.7)
#667 Claude 2 (43.7)