GPT-4 (0613)
OpenAI's flagship model, GPT-4 is a large-scale multimodal language model capable of solving difficult problems with greater accuracy than previous models due to its broader general knowledge and advanced reasoning.
text → text · made by OpenAI
Sold by 4 ways
| Seller | Lane | Rate |
|---|---|---|
| OpenAI api | standard | $30per Mtok in$60per Mtok outdevelopers.openai.com · read 2026-08-24 |
| Microsoft Azure AI cloud | standard | $60per Mtok in$120per Mtok outraw.githubusercontent.com · read 2026-09-16 |
| Nous Research aggregator | standard | $30per Mtok in$60per Mtok outinference-api.nousresearch.com · read 2026-09-12 |
| OpenRouter aggregator | standard | $30per Mtok in$60per Mtok outopenrouter.ai · read 2026-08-24 |
Measured 19 standings
| Place | Board | Metric | Score |
|---|---|---|---|
| 1stof 9 | AlpacaEval Leaderboard | Win Rate | 95.3 |
| 7thof 170 | Gsm8k — Epoch AI | EM | 0.9 |
| 11thof 115 | Trivia qa — Epoch AI | EM | 0.848 |
| 13thof 88 | Bbh — Epoch AI | Average | 0.751 |
| 20thof 217 | Mmlu — Epoch AI | EM | 0.824 |
| 22ndof 47 | WebArena Leaderboard (official Google Sheet) | Success Rate (%) | 23.5 |
| 45thof 50 | Metr time horizons — Epoch AI | average_score | 0.293 |
| 59thof 81 | Forecastbench — Epoch AI | Overall score | 57.8 |
| 85thof 108 | Math level 5 — Epoch AI | mean_score | 22.97 |
| 92ndof 127 | Mystery game puzzles — Epoch AI | mean_score | 12 |
| 95thof 123 | Lmca — Epoch AI | Score | 17.06 |
| 105thof 161 | Dtbench — Epoch AI | Accuracy | 62.67 |
| 161stof 170 | Weirdml — Epoch AI | Accuracy | 12.36 |
| 167thof 184 | Artificial Analysis · Coding Index | Index | 13.1 |
| 169thof 222 | Chess puzzles — Epoch AI | mean_score | 4 |
| 281stof 290 | OTIS Mock AIME 2024-2025 | mean_score | 1.111 |
| 288thof 313 | GPQA Diamond | mean_score | 30.65 |
| 293rdof 402 | LMArena · Text | Score (Elo) | 1276 |
| 562ndof 656 | Epoch capabilities index — Epoch AI | ECI Score | 123.09 |
About
GPT-4 (0613) — a text model from OpenAI, sold by 4 companies from $30 in and $60 out per million tokens, placed 1st of 9 on AlpacaEval Leaderboard.
It takes text and returns text, with a context window of 8,191 tokens. It was published in November 2023, trained on material up to November 2023. Its sellers say it can call a tool. The catalogue files it under reasoning and agents. Four companies sell it. The cheapest is $30 in and $60 out per million tokens at Nous Research. It has been measured on 19 boards, and stands best at 1st of 9 on AlpacaEval Leaderboard.
Every current figure
- Maker
- OpenAI
- Register
- model
- Takes
- text
- Returns
- text
- Context
- 8,191 tokens
- Published
- November 2023
- Knowledge to
- November 2023
- Licence
- not read
- Sellers
- 4
- Price
- $30 in and $60 out per million tokens — Nous Research
- Boards
- 19
- Best place
- 1st of 9 — AlpacaEval Leaderboard
Known as 3 names
BrowserGym + GPT-4openai/gpt-4gpt-4-0613