LLM Comparison
Mistral Small 4 vs GPT-5.5
Side-by-side specs, pricing & capabilities · Updated May 2026
Price vs Intelligence
Add to comparison
2/6 modelsSame tier:
| Organization | ||
| OpenTools Score | 21 55.7 | 90 5.1 |
| Family | Mistral | GPT |
| Status | Current | Current |
| Release Date | Mar 2026 | Apr 2026 |
| Context Window | 262K tokens | 1.1M tokens |
| Input Price | $0.15/M tokens | $5.00/M tokens |
| Output Price | $0.60/M tokens | $30.00/M tokens |
| Pricing Notes | Cache read: $0.0150/M tokens | Cached input: $0.50/M tokens. Long context (>272K tokens): 2x input, 1.5x output. Batch API: 50% discount. Priority: 2.5x standard. |
| Capabilities | textvisioncodetool-use | textvisioncodetool-useextended-thinkingcomputer-useweb-search |
| Training Cutoff | — | December 2025 |
| Max Output | — | 128K tokens |
| API Identifier | mistralai/mistral-small-2603 | openai/gpt-5.5 |
| Benchmarks | ||
| MMLU-Pro | 68.5mistral | — |
| HumanEval | 78.5mistral | — |
| MGSM | 72.3mistral | — |
| GPQA Diamond | 58.8mistral | 93.6openai |
| LiveCodeBench | 64mistral | — |
| MMLU | — | 92.4openai |
| ARC-AGI-2 | — | 85openai |
| Terminal-Bench 2.0 | — | 82.7openai |
| SWE-bench Pro | — | 58.6openai |
| OSWorld-Verified | — | 78.7openai |
| BrowseComp | — | 84.4openai |
| MMMU-Pro | — | 81.2openai |
| FrontierMath Tier 4 | — | 35.4openai |
| HLE (with tools) | — | 52.2openai |
| GDPval | — | 84.9openai |
| Toolathlon | — | 55.6openai |
| CyberGym | — | 81.8openai |
| MRCR v2 512K-1M | — | 74openai |
| View Mistral Small 4 | View GPT-5.5 | |
Cost Calculator
Enter your expected monthly token usage to compare costs.
| Model | Input | Output | Total / mo | vs Best |
|---|---|---|---|---|
| Mistral Small 4Cheapest | $0.15 | $0.30 | $0.45 | — |
| GPT-5.5 | $5.00 | $15.00 | $20.00 | +4344% |
Mistral AI
Mistral Small 4
Mistral Small 4 is a multimodal llm from Mistral AI. Supports up to 262,144 token context window. Achieves 76.5% on MMLU. Available from $0.15/M input tokens.
OpenAI
GPT-5.5
GPT-5.5 is OpenAI's smartest and most intuitive model, built for agentic work like coding, research, and data analysis. It matches GPT-5.4 per-token latency while delivering higher intelligence with significantly fewer tokens. Supports a 1,050,000 token context window and five reasoning effort levels (none through xhigh).
More Comparisons
Looking for more AI models?
Browse All LLMs