Welcome to the first edition of our monthly LLM rankings. Each month, we pull real-world usage data from OpenRouter — a platform routing millions of API calls across every major LLM — and rank the models developers are actually paying for.
Not benchmarks. Not press releases. Real tokens, real money, real usage.
The Top 10 — January 2026
| Rank | Model | Provider | Move | Input / 1K tokens | Output / 1K tokens |
|---|---|---|---|---|---|
| 1 | Claude 3.5 Sonnet | Anthropic | ● | $0.003 | $0.015 |
| 2 | Gemini 2.0 Flash | ↑ | $0.0001 | $0.0004 | |
| 3 | GPT-4o | OpenAI | ↓ | $0.005 | $0.015 |
| 4 | DeepSeek V3 | DeepSeek | ↑ | $0.0003 | $0.0009 |
| 5 | Gemini 2.0 Pro | ● | $0.002 | $0.008 | |
| 6 | Claude 3 Opus | Anthropic | ↓ | $0.015 | $0.075 |
| 7 | GPT-5.2 | OpenAI | 🆕 | $0.00175 | $0.014 |
| 8 | Devstral 2 | Mistral | 🆕 | $0.0004 | $0.002 |
| 9 | DeepSeek R1 | DeepSeek | ↑ | $0.0006 | $0.002 |
| 10 | Llama 3.3 70B | Meta | ↓ | $0.0001 | $0.0004 |
New This Month
GPT-5.2 (Dec 10) — OpenAI's major winter release. 400K context window, significantly better reasoning than 4o, and a more aggressive price point. Early adoption is strong but hasn't yet dethroned Claude for coding workflows.
Devstral 2 (Dec 9) — Mistral's developer-focused model. 262K context, competitive pricing, and strong code generation. A serious entry in the mid-tier coding model space.
Gemini 3 Flash Preview (Dec 17) — Google's next-gen speed model drops in late December. Early access users report absurd throughput numbers. Expect this one to climb fast in February.
The Story
Claude 3.5 Sonnet starts 2026 exactly where it ended 2025: on top. Developers have built entire toolchains around it, and switching costs are real. But the challengers are circling. DeepSeek V3 offers 90% of the quality at 10% of the price. GPT-5.2 brings OpenAI back into the conversation with a genuinely improved model. And Google's Flash lineup is eating the high-volume, low-cost segment alive.
Market Share
| Provider | Share | Trend |
|---|---|---|
| Anthropic | ~30% | Stable |
| ~25% | Growing | |
| OpenAI | ~20% | Recovering |
| DeepSeek | ~12% | Rising fast |
| Meta | ~6% | Stable |
| Mistral | ~4% | Growing |
| Others | ~3% | — |
Monthly LLM rankings by Demand Signals, sourced from OpenRouter usage data. Subscribe to our blog for monthly updates.
Get a Free AI Demand Gen Audit
We'll analyze your current visibility across Google, AI assistants, and local directories — and show you exactly where the gaps are.