Compact GPT model for low-latency assistance and high-volume workloads
Specification
most-agreed values
Context
400K
Max output
128K
Released
2026-05-05
Knowledge cutoff
2025-12-01
Retires
—
Open weights
no
Input
text, image, pdf
Output
text
Price
US dollars per million tokens · most-agreed
Input
$5.00
Output
$30.00
Cache read
$0.50
Quality
5 benchmarks · 1 source
Composite
48th
percentile of 334 scored models
Rank
173 / 334
±0.08 sd
Evidence
5 × 1
one source, or fewer than 3 benchmarks
Effort range
—
one configuration only
reasoning73rd2/4 bench
math36th3/7 bench
Every score
one row per source and configuration — nothing averaged away
| Benchmark | Score | Configuration | Source | Run |
|---|---|---|---|---|
| GPQA diamond | 82.5 ±2.3 | default | Epoch AI | 2026-08-02 |
| OTIS Mock AIME 2024-2025 | 68.1 ±6.2 | default | Epoch AI | 2026-08-02 |
| Chess Puzzles | 12.0 ±3.3 | default | Epoch AI | 2026-08-02 |
| FrontierMath-Tiers-1-3-v2 | 26.3 ±2.6 | default | Epoch AI | 2026-08-02 |
| FrontierMath-Tier-4-v2 | 2.4 ±2.4 | default | Epoch AI | 2026-08-02 |
The leading figure for a benchmark is the median across its configurations, so one heroic high-effort run cannot set the number.
Available from 1 host
| Host | In $/M | Out $/M | Cache rd | Cache wr | Context | Output | Retires |
|---|---|---|---|---|---|---|---|
| zenmux | $5.00 | $30.00 | $0.50 | — | 400K | 128K | — |
More from OpenAI
most-hosted first