Tier
Major
Headquarters
San Francisco, CA
Founded
2015
Flagship
GPT
Models tracked
86
Artificial Analysis
Top intelligence
58.9
Artificial Analysis Index
Cheapest model
$0.00
blended /1M
Fastest output
257
tokens / sec
Flagship Model
OpenAI’s top-scoring model on the Artificial Analysis Intelligence Index.
Most intelligent model
GPT-5.6 Sol (max)
Released 2026-07-09
Intelligence
58.9
Coding
77.4
Math
—
In $/1M
$5.00
Out $/1M
$30.00
Speed t/s
70
Benchmark breakdown
Models & benchmarks
Every OpenAI model on file — intelligence, coding, price, and speed.
| Model | Released | Intel | Coding | Math | In $/1M | Out $/1M | Speed t/s | Latency s |
|---|---|---|---|---|---|---|---|---|
| GPT-5.6 Sol (max) | 2026-07-09 | 58.9 | 77.4 | — | $5.00 | $30.00 | 70 | 56.83 |
| GPT-5.6 Sol (xhigh) | 2026-07-09 | 57.7 | 78.3 | — | $5.00 | $30.00 | 74 | 16.82 |
| GPT-5.6 Sol (high) | 2026-07-09 | 55.9 | 77.2 | — | $5.00 | $30.00 | 67 | 8.90 |
| GPT-5.6 Terra (max) | 2026-07-09 | 55.0 | 76.7 | — | $2.50 | $15.00 | 150 | 113.37 |
| GPT-5.5 (xhigh) | 2026-04-23 | 54.8 | 74.9 | — | $5.00 | $30.00 | 0 | 0.00 |
| GPT-5.6 Sol (medium) | 2026-07-09 | 53.6 | 76.3 | — | $5.00 | $30.00 | 66 | 6.12 |
| GPT-5.5 (high) | 2026-04-23 | 53.1 | 71.6 | — | $5.00 | $30.00 | 0 | 0.00 |
| GPT-5.6 Terra (xhigh) | 2026-07-09 | 51.6 | 70.6 | — | $2.50 | $15.00 | 122 | 6.87 |
| GPT-5.4 (xhigh) | 2026-03-05 | 51.4 | 71.1 | — | $2.50 | $15.00 | 0 | 0.00 |
| GPT-5.6 Luna (max) | 2026-07-09 | 51.2 | 71.4 | — | $1.00 | $6.00 | 204 | 73.70 |
| GPT-5.5 (medium) | 2026-04-23 | 50.4 | 71.5 | — | $5.00 | $30.00 | 0 | 0.00 |
| GPT-5.6 Sol (low) | 2026-07-09 | 49.4 | 69.7 | — | $5.00 | $30.00 | 61 | 2.62 |
| GPT-5.6 Luna (xhigh) | 2026-07-09 | 49.1 | 68.6 | — | $1.00 | $6.00 | 188 | 17.67 |
| GPT-5.6 Terra (high) | 2026-07-09 | 49.0 | 67.1 | — | $2.50 | $15.00 | 121 | 2.30 |
Trajectory
OpenAI’s model releases over time, by intelligence.
House strengths
Where OpenAI’s lineup over-indexes, averaged across its models.
Capability profile
Best value
gpt-oss-20b (high)
Most intelligence per dollar — 14.9 index at $0.10/1M blended.
157 index points per $1
Governance
OpenAI’s license posture and openness across its lineup.
Open weight
4
of 51 models
Proprietary
47
of 51 models
Commercial-OK
4
license permits
Modalities
Image · Text
Via LLM Stats
Media models
Image, video, and speech models, ranked by arena Elo.
| Model | Type | Released | Elo | Rank |
|---|---|---|---|---|
| GPT Image 2 (high) | text to image | 2026-04 | 1340 | #1 |
| GPT Image 1.5 (high) | text to image | 2025-12 | 1264 | #5 |
| GPT Image 2 (high) | image editing | 2026-04 | 1259 | #1 |
| GPT Image 1.5 (high) | image editing | 2025-12 | 1255 | #2 |
| Sora 2 Pro | text to video | 2025-09 | 1181 | — |
| Sora 2 (December) | text to video | 2025-12 | 1169 | — |
| Sora 2 (October) | text to video | 2025-09 | 1162 | — |
| GPT Image 1 (high) | text to image | 2025-04 | 1139 | #49 |
| GPT Image 1 (high) | image editing | 2025-04 | 1135 | #45 |
| TTS-1 HD | text to speech | — | 1102 | #26 |
| GPT Image 1 Mini (medium) | text to image | 2025-10 | 1094 | #77 |
| TTS-1 | text to speech | — | 1083 | #33 |
Publications
OpenAI’s own papers, frameworks, and cards in the 44B Library.
Report · 2026
Accelerating scientific discovery with ChatGPT for Academic Researchers
Report · 2026
How GPT-5.6 fuses frontier intelligence with frontier efficiency
Report · 2026
How enabling two settings tripled our scores on the ARC-AGI-3 benchmark
Report · 2026
Scientific computing in the age of agentic AI
Report · 2026
How AI is expanding what people do at work
Report · 2026
Launching Health in ChatGPT
Report · 2026
Advancing the next era of national science
Report · 2026
Building AI infrastructure with the Effingham County community
Report · 2026
How news organizations are using AI to advance their vital missions
Report · 2026
Introducing OpenAI Presence
Report · 2026
NTT DATA Group cuts incident analysis to 30 minutes with Codex
Report · 2026
David Vélez and Robin Vince join the boards of the OpenAI Foundation and OpenAI Group PBC