Models Directory.
Compare raw specifications, developer benchmarks, context caps, and token costs of modern AI backbones.
| Model Name | Creator | Type | API Cost (In/Out) | Context | MMLU | Coding | Reasoning | Privacy | Release Date | Actions |
|---|---|---|---|---|---|---|---|---|---|---|
| Claude 4 Opus LatestPHD Reasoning | Anthropic | Proprietary | $15.00 / $75.00M | 500k | 94.8% | 97.5% | 96.8% | High | 2026-04-18 | |
| GPT-5.5 Pro LatestMath Logic | OpenAI | Proprietary | $10.00 / $40.00M | 256k | 94.6% | 96.2% | 97.4% | Medium | 2026-03-11 | |
| DeepSeek V4 Pro LatestBest Value | DeepSeek | Open Weights | $0.20 / $0.80M | 256k | 93.2% | 94% | 95.2% | Low | 2026-05-02 | |
| Gemini 3.1 Pro 4M ContextResearch | Proprietary | $2.50 / $10.00M | 4M | 92.5% | 91% | 93.8% | Medium | 2026-02-15 | ||
| Llama 4 Scout 10M ContextSelf-Hostable | Meta | Open Weights | $0.20 / $0.60M | 10M | 89.2% | 87.5% | 88% | High | 2026-05-19 | |
| Claude 3.7 Sonnet CodingAgentic | Anthropic | Proprietary | $3.00 / $15.00M | 200k | 91.2% | 95.8% | 93.4% | High | 2025-02-25 | |
| Gemini 2.0 Flash SpeedMultimodal | Proprietary | $0.07 / $0.30M | 1M | 82.5% | 83.2% | 84% | Medium | 2024-12-11 |
Claude 4 Opus
Anthropic • Released 2026-04-18
Anthropic's flagship 2026 model. Delivers PHD-level scientific reasoning, peerless multi-file codebase operations, and fully autonomous agent orchestration.
GPT-5.5 Pro
OpenAI • Released 2026-03-11
OpenAI's latest generation model. Features dynamic reasoning pipelines and incredible native visual-spatial comprehension.
DeepSeek V4 Pro
DeepSeek • Released 2026-05-02
DeepSeek's flagship open-weights champion. Matches proprietary models on key benchmarks at 1/50th of their API price ratios.
Gemini 3.1 Pro
Google • Released 2026-02-15
Google's 2026 powerhouse. Features a massive 4 million token context window and native video/audio document comprehension.
Llama 4 Scout
Meta • Released 2026-05-19
Meta's context-focused openweights release. Highlights a massive 10 million token context window, designed for complete local database uploads.
Claude 3.7 Sonnet
Anthropic • Released 2025-02-25
Anthropic's stable reasoning flagship. Combines hybrid coding capability with clean markdown outputs.
Gemini 2.0 Flash
Google • Released 2024-12-11
Google's high speed model. Features sub-second latency and 1M token context capacity.