Small GPT-5 for responsive agents, coding help, and everyday automation
Specification
openai listing
Context
400K
Max output
128K
Released
2025-08-07
Knowledge cutoff
2024-05-30
Retires
2026-12-11
Open weights
no
Input
text, image
Output
text
Price
US dollars per million tokens · openai list price
Input
$0.25
Output
$2.00
Cache read
$0.025
Cache write
$0.25
Batch and priority prices
US dollars per million tokens · per listing
| Price | Listing | $/M |
|---|---|---|
| Batch input | openai | $0.125 |
| Batch output | openai | $1.00 |
| Priority input | openai | $0.45 |
| Priority output | openai | $3.60 |
Batch input is 50% below the interactive input price at the same listing.
Quality
11 benchmarks · 2 sources
Composite
54th
percentile of 334 scored models
Rank
153 / 334
±0.05 sd
Evidence
11 × 2
3 or more benchmarks, from 2 or more sources
Effort range
3.3
points between effort settings
coding48th1/3 bench
reasoning54th3/4 bench
math62nd6/7 bench
knowledge18th1/1 bench
By reasoning effort
every setting placed on the same scale as the leaderboard
| Setting | Composite | Benchmarks behind it |
|---|---|---|
| high | 61st | 10 |
| medium | 57th | 9 |
Every score
one row per source and configuration — nothing averaged away
| Benchmark | Score | Configuration | Source | Run |
|---|---|---|---|---|
| MATH level 5 | 97.8 ±0.3 | high | Epoch AI | 2025-10-30 |
| ↳ | 96.8 ±0.4 | medium | Epoch AI | 2025-08-20 |
| Settings differ by 1.0 points on this benchmark. | ||||
| OTIS Mock AIME 2024-2025 | 86.7 ±4.3 | high | Epoch AI | 2025-10-30 |
| ↳ | 66.9 ±7.5 | medium · 2 runs | Epoch AI | 2026-08-07 |
| Settings differ by 19.8 points on this benchmark. | ||||
| GPQA diamond | 75.0 ±2.3 | high | Epoch AI | 2025-10-30 |
| ↳ | 71.7 ±3.2 | medium · 2 runs | Epoch AI | 2026-08-07 |
| Settings differ by 3.3 points on this benchmark. | ||||
| FrontierMath-v1 | 30.2 ±2.4 | medium · 2 runs | Epoch AI | 2025-11-13 |
| ↳ | 28.6 ±2.6 | high · 2 runs | Epoch AI | 2025-11-13 |
| Settings differ by 1.6 points on this benchmark. | ||||
| Chess Puzzles | 30.0 ±4.6 | high | Epoch AI | 2026-08-07 |
| ↳ | 9.5 ±2.6 | medium · 2 runs | Epoch AI | 2026-08-07 |
| Settings differ by 20.5 points on this benchmark. | ||||
| SWE-Bench verified | 64.7 ±2.2 | medium | Epoch AI | 2026-02-01 |
| ↳ | 59.8 | medium · mini-SWE-agent | SWE-bench | — |
| ↳ | 56.2 | default · mini-SWE-agent | SWE-bench | — |
| Sources and settings disagree by 8.5 points. Both figures stand. | ||||
| FrontierMath-Tier-4-2025-07-01 | 3.1 ±0.0 | high · 2 runs | Epoch AI | 2025-10-30 |
| ↳ | 2.1 ±0.0 | medium · 2 runs | Epoch AI | 2025-08-07 |
| Settings differ by 1.0 points on this benchmark. | ||||
| FrontierMath-Tiers-1-3-v2 | 46.7 ±3.0 | high | Epoch AI | 2026-06-12 |
| ↳ | 12.1 ±1.4 | medium · 2 runs | Epoch AI | 2026-08-27 |
| Settings differ by 34.6 points on this benchmark. | ||||
| FrontierMath-Tier-4-v2 | 12.2 ±5.2 | high | Epoch AI | 2026-06-12 |
| SimpleQA Verified | 21.6 ±1.3 | high | Epoch AI | 2026-08-10 |
| Mystery Game Puzzles | 7.0 ±2.0 | medium · 2 runs | Epoch AI | 2026-08-27 |
| ↳ | 5.0 ±2.2 | high | Epoch AI | 2026-08-27 |
| Settings differ by 2.0 points on this benchmark. | ||||
The leading figure for a benchmark is the median across its configurations, so one heroic high-effort run cannot set the number.
Available from 36 hosts
| Host | In $/M | Out $/M | Cache rd | Cache wr | Context | Output | Retires |
|---|---|---|---|---|---|---|---|
| openaivendor | $0.25 | $2.00 | $0.025 | — | 400K 272K | 128K | 2026-12-11 |
| qihang-ai | $0.04 | $0.29 | — | — | 200K | 64K | — |
| poe | $0.22 | $1.80 | $0.022 | — | 400K | 128K | — |
| jiekou | $0.225 | $1.80 | — | — | 400K | 128K | — |
| 302ai | $0.25 | $2.00 | — | — | 400K | 128K | — |
| abacus | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| azure | $0.25 $0.275 | $2.00 $2.20 | $0.03 $0.0275 $0.025 | — | 400K 272K | 128K | 2027-02-09 |
| azure-cognitive-services | $0.25 | $2.00 | $0.03 | — | 400K | 128K | — |
| cloudflare-ai-gateway | $0.25 | $2.00 | $0.025 | — | 128K | 128K | — |
| databricks | $0.25 | $2.00 | $0.025 | $0.25 | 400K 272K | 128K | — |
| edenai | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| fastrouter | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| github-copilot | $0.25 | $2.00 | $0.025 | — | 264K | 64K | — |
| helicone | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| impossibl | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| kilo | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| llmgateway | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| llmgateway-providers | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| merge-gateway | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| nano-gpt | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| nearai | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| neon | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| oci · oci | $0.25 | $2.00 | — | — | 272K | 128K | — |
| ofox | $0.25 | $2.00 | $0.03 | — | 256K | 33K | — |
| openrouter | $0.25 | $2.00 | $0.025 | — | 400K 272K | 128K | — |
| orcarouter | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| perplexity | $0.25 | $2.00 | $0.025 | — | — | — | — |
| perplexity-agent | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| pioneer | $0.25 | $2.00 | $0.025 | $0.25 | 400K | 128K | — |
| replicate · replicate | $0.25 | $2.00 | — | — | — | — | — |
| sap-ai-core | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| vercel | $0.25 | $2.00 | $0.025 | — | 400K | 128K | — |
| vivgrid | $0.25 | $2.00 | $0.03 | — | 272K | 128K | — |
| cortecs | $0.279 | $2.192 | $0.056 | — | 400K | 128K | — |
| anyapi | — | — | — | — | 400K | 128K | — |
| github_copilot · github-copilot | — | — | — | — | 128K | 64K | — |
More from OpenAI
most-hosted first