GPT-3.5: benchmark scores, price and ranking

GPT-3.5 from OpenAI ranks #690 of 743 LLMs on the LLMs Tiger aggregate benchmark score (43.0), based on 3 benchmark results captured up to 2026-10-13.

Organisation: OpenAI
Type: Frontier model
Open weights: No
Released: 2022-11-30 (source: curated)
Aggregate rank: #690 of 743 (score 43.0)

Relative to other models, GPT-3.5 is strongest on SWE-bench Verified (#73 of 73) and weakest on SWE-bench Test (#11 of 11).

Across the 3 benchmarks where at least ten models are scored, GPT-3.5 is in the top 10% on 0 and in the top half on 0.

GPT-3.5 benchmark results

GPT-3.5 benchmark scores, ranked against all models measured on each benchmark
Benchmark	Score	Rank	Source	Captured
SWE-bench Verified	0.4%	#73 of 73	SWE-bench Verified leaderboard	2026-10-13
SWE-bench Lite	0.3%	#20 of 20	SWE-bench Lite leaderboard	2026-10-13
SWE-bench Test	0.2%	#11 of 11	SWE-bench Test leaderboard	2026-10-13

About GPT-3.5

GPT-3.5 is a frontier model from OpenAI. It was released on 2022-11-30, according to the developer's announcement. No API price is listed for it yet. We hold 3 benchmark results for it across 2 areas, and none is in the top tenth of its board.

Where GPT-3.5 stands

On our aggregate score, GPT-3.5 is #690 of 743 ranked models, in the lower half. Its score is 43.0; the scale is centred on 50, so it sits 7.0 points below the typical model. That beats 7% of the ranked field. Among the 88 ranked OpenAI models it is #81. The next one up is GPT-4.1 mini at 43.0. The next one down is gpt-4-turbo-2024-04-09 at 42.6.

GPT-3.5, benchmark by benchmark

SWE-bench Verified is a weak spot. GPT-3.5 scores 0.4%, #73 of 73, behind 99% of the models measured. The leader, Claude Opus 4.7 (max), reaches 83.5%. It covers real GitHub issues resolved. The board median is 56.4%, so GPT-3.5 is 56.0 points below it. Source: SWE-bench Verified leaderboard, captured 2026-10-13.

SWE-bench Lite is a weak spot. GPT-3.5 scores 0.3%, #20 of 20, behind 95% of the models measured. The leader, Claude 4 Sonnet, reaches 56.7%. It covers real GitHub issues resolved. The board median is 24.7%, so GPT-3.5 is 24.3 points below it. Source: SWE-bench Lite leaderboard, captured 2026-10-13.

SWE-bench Test is a weak spot. GPT-3.5 scores 0.2%, #11 of 11, behind 91% of the models measured. The leader, Claude 4.5 Opus, reaches 52.6%. It covers real GitHub issues resolved. The board median is 12.5%, so GPT-3.5 is 12.3 points below it. Source: SWE-bench Test leaderboard, captured 2026-10-13.

Strengths and weaknesses

Software engineering: weak, ahead of 0% of the field on average across 2 benchmarks. Its best board here is SWE-bench Lite (#20 of 20) and its weakest is SWE-bench Lite (#20 of 20).

Release history

The next one was GPT-4 (Mar 2023) on 2023-03-14, at 43.7.

Reading these numbers

The 3 scores on this page come from 3 sources: SWE-bench Lite leaderboard (1), SWE-bench Test leaderboard (1), SWE-bench Verified leaderboard (1). 3 of them come from an official leaderboard and 0 from datasets or third-party runs. With 3 results the aggregate rank can move a lot when one new score arrives.

Models ranked nearby

#687 claude-3-opus-20240229 (43.2)
#688 granite-3.1-3b-a800m-base (43.1)
#689 GPT-4.1 mini (43.0)
#691 gemma-2-9b-it (43.0)
#692 Llama-2-70b-chat-hf (42.9)
#693 SmolLM2-1.7B (42.9)