GPT-4

GPT-4 is a model from OpenAI. 1 benchmark counts toward its score, in 1 category.

partialShows if the model has enough results for an index.
IndexOverall score out of 100.Unranked
CoverageShare of the index weight with results.20%
SpeedOutput tokens per second.44/s
Input / 1MUS dollars per 1M input tokens.$30
Output / 1MUS dollars per 1M output tokens.$60
ContextMaximum tokens in one request.8K
EloLMArena rating and rank.1186 (#314)

The index is a score out of 100. The ± range shows how much it can change.

88,723 votes. Elo shows what people prefer. It does not change the score.

CapabilitiesScore per category, out of 100.

Out of 100
AgenticMulti-step tasks with tools.
N/A
CodingCode writing and repair.
21.4
ReasoningLogic problems and puzzles.
N/A
MultimodalTasks with images and text.
N/A
KnowledgeFacts and expert knowledge.
N/A
MultilingualTasks in many languages.
N/A
InstructionTasks with strict rules in the prompt.
N/A
MathMath problems.
N/A

Results

1 counted
BenchmarkThe test name.CategoryThe capability that the test measures.ResultThe score from the publisher.IndexThis result as a score out of 100.RunThe settings of the run.DateDate of the result.Published byThe source of the result.
SWE-bench LiteCoding23.7%23.4OpenCSG StarShip CodeGenAgent11 Sept 2025SWE-bench team
SWE-bench LiteCoding19.0%19.4AutoCodeRover11 Sept 2025SWE-bench team

1 benchmarks count, from 2 of 2 results. A grey row does not count. Too few models took that benchmark.

Sources

BenchLM benchmark aggregationUsed with attribution; per-benchmark results credited to their original authorsOpenRouter, collected directlyNo licence statedSWE-bench team, collected directlyNo licence stated. The repository publishes submission records for reproducibility and transparency and asks that SWE-bench be cited

More from OpenAI

GPT-6 Astra83.5GPT-5.6 Sol76.7GPT-5.6 Terra73.4GPT-6 Sol76.9GPT-5.572.3