GPT-5.5
OpenAIProprietaryReasoningVerifiedReleased Apr 23, 2026Input: file, image, text
Overall score
69.2
Rank 20 of 222 · 63.0–75.5
Context window
1M
1,000,000 tokens
Price per 1M tokens
$5.00 in · $30.00 out
✓ Verified on OpenAI's pricing page
Blended price
$11.25
3:1 input:output tokens
Output speed
—
Not measured yet
Category scores
Scores per capability area with the model's rank among ranked models in that category.
- 1Instruction following#893.3
- 2Math91.2
- 3Agentic & tool use#2381.7
- 4Knowledge#1580.8
- 5Multimodal#2375.5
- 6Reasoning#1770.3
- 7Coding#1455.5
Benchmark results (40)
As reported in system cards, papers and public evaluations. Position compares this model with every other model reporting the same benchmark.
Agentic & tool use
| Gert Labs | 72.93 | #2 of 52 |
| BrowseComp | 84.4 | #14 of 50 |
| MCP Atlas | 75.3 | #18 of 42 |
| MazeBench | 0 | #20 of 42 |
| Terminal-Bench 2.0 | 82.0 | #1 of 39 |
| OSWorld-Verified | 78.7 | #12 of 35 |
| JobBench | 42.7 | #16 of 34 |
| CyberGym | 81.8 | #12 of 25 |
| OSWorld 2.0 | 13.0 | #20 of 24 |
| Toolathlon | 55.6 | #5 of 22 |
| ResearchClawBench | 17.0 | #11 of 19 |
| ApprenticeBench | 20 | #7 of 18 |
| ExploitGym | 13.4 | #9 of 14 |
Multimodal
| Design Arena Website | 1257 | #40 of 91 |
| MMMU-Pro | 81.2 | #7 of 48 |
| ParseBench | 67.76 | #19 of 37 |
| ExtractBench (LlamaIndex)(lower is better) | 93.57 | #21 of 24 |
| OfficeQA Pro | 54.1 | #10 of 13 |
| MMMU-Pro w/ Python | 83.2 | #3 of 9 |
Reasoning
| ARC-AGI-2 | 85.0 | #11 of 41 |
| ARC-AGI-1 | 95.00 | #13 of 36 |
| ARC-AGI-3 | 0.4 | #14 of 22 |
| MRCR v2 64K-128K | 83.1 | #2 of 2 |
| MRCR v2 128K-256K | 87.5 | #1 of 2 |
Coding
| SWE-bench Pro | 58.6 | #31 of 69 |
| FrontierCode 1.1 Main | 43.0 | #10 of 16 |
| React Native Evals | 84.7 | #3 of 15 |
| PostTrainBench v1.1 | 27.2 | #12 of 15 |
Knowledge
| GPQA | 93.6 | #5 of 117 |
| HLE | 52.2 | #15 of 69 |
| HLE w/o tools | 41.4 | #21 of 43 |
Math
| FrontierMath v2 (Tiers 1-3) | 85.263 | #12 of 64 |
| FrontierMath v2 (Tier 4) | 72.500 | #12 of 51 |
| FrontierMath (legacy) | 51.7 | #5 of 7 |
Pricing by source
USD per 1M tokens.
| Source | Input | Output | Cached in |
|---|---|---|---|
| ✓ Official (OpenAI) checked Oct 17, 2026 · gpt-5.5 | $5.00 | $30.00 | $0.50 |
| BenchLM list price | $5.00 | $30.00 | $0.50 |
| OpenRouter openai/gpt-5.5 | $5.00 | $30.00 | $0.50 |
Standard tier, prompts under 272K tokens.
OpenAI's current API pricing page lists gpt-5.5 at $5.00 input / $0.50 cached input / $30.00 output per million tokens for short-context requests. Batch and Flex are available at half the standard rate, and Priority is 2.5x the standard rate.
Arena ratings
Human-preference ratings from LMArena. Best-rated variant shown.
| Text(high) | 1471 | #28 · 69.4K votes |
| WebDev (coding)(xHigh, Codex Harness) | 1513 | #58 · 15.6K votes |
| Vision | 1297 | #21 · 28.8K votes |
| Documents | 1486 | #8 · 21.3K votes |
Speed
Measured by us through OpenRouter (median of trials).
Not measured yet.
Related models
Same family
- GPT-5.5 Pro73.5
Similar score
- Gemini 3.8 Flash69.2 · $1.50
- Step 5 Preview69.6 · $1.43
- GPT-5.4 Pro68.7 · $67.50
- Claude Opus 4.7 (Adaptive)68.7 · $10.00
- Qwen3.8 Max69.7 · —