LLM Token Counter & Monthly API Cost Calculator
Live pricing ledger synchronized with OpenRouter v1 API. Filter by provider, search across 465+ models, and model Prompt Caching hit ratios plus 50% Batch API discounts.
OpenRouter v1 Registry (24 curated 2026 models · 465 upstream)
Tokens per call
4,800
Lowest cost 2026 model
DeepSeek V4.1 Flash 1017
Lowest 30-day bill
$12.66/mo
2026 models compared
24
| Model (2026 Release) | Provider & date | Input / cached ($/1M) | Output ($/1M) | Per 1K calls | 30-day spend |
|---|---|---|---|---|---|
| Gemini Omni 1.2 Flash | Google 2026-11-05 | $1.5 / $0.375 | $9 | $10.9500 | |
| GPT-6.1 Terra Pro | OpenAI 2026-10-26 | $1 / $0.05 | $5 | $6.1000 | |
| Kimi K3.1 Pro | Moonshot 2026-10-23 | $2 / $0.2 | $10 | $12.4000 | |
| MiniMax M3.1 Flash | MiniMax 2026-10-19 | $0.15 / $0.015 | $0.6 | $0.8100 | |
| DeepSeek V4.1 Flash 1017 | DeepSeek 2026-10-17 | $0.006 / $0.005 | $0.5 | $0.4220 | |
| Gemini 4 Argon | Google 2026-10-13 | $2 / $0.1 | $10 | $12.2000 | |
| Claude Haiku 5.5 | Anthropic 2026-10-07 | $0.1 / $0.01 | $0.5 | $0.6200 | |
| Mistral Large 4 | Mistral 2026-10-06 | $0.7 / $0.07 | $2.1 | $3.2200 | |
| GPT-6.1 Sol Pro | OpenAI 2026-09-29 | $2 / $0.1 | $10 | $12.2000 | |
| GPT-6.1 Sol | OpenAI 2026-09-29 | $2 / $0.1 | $10 | $12.2000 | |
| Claude Sonnet 5.5 | Anthropic 2026-09-28 | $2 / $0.1 | $10 | $12.2000 | |
| GLM 5.3 Prime | Z.ai / GLM 2026-09-23 | $2.8 / $0.56 | $8.8 | $13.7600 | |
| Qwen3.8 Max Prime | Qwen 2026-09-23 | $4 / $0.5 | $12 | $18.6000 | |
| GPT-6 Luna | OpenAI 2026-09-22 | $0.1 / $0.01 | $0.5 | $0.6200 | |
| Claude Opus 5.5 | Anthropic 2026-09-22 | $4 / $0.2 | $20 | $24.4000 | |
| Grok 4.7 | xAI 2026-09-21 | $2 / $0.5 | $6 | $9.8000 | |
| Qwen3.8 Omni Flash | Qwen 2026-09-21 | $0.15 / $0.016 | $0.47 | $0.7080 | |
| DeepSeek V4.1 Flash | DeepSeek 2026-09-10 | $0.006 / $0.005 | $0.5 | $0.4220 | |
| GPT-6 Astra Pro | OpenAI 2026-09-04 | $10 / $1 | $50 | $62.0000 | |
| Gemini 3.8 Flash | Google 2026-09-02 | $0.825 / $0.083 | $4.125 | $5.1150 | |
| Claude Fable 5.1 | Anthropic 2026-09-01 | $10 / $0.25 | $50 | $60.5000 | |
| GLM 5.3 Flash | Z.ai / GLM 2026-08-26 | $0.15 / $0.03 | $0.5 | $0.7600 | |
| DeepSeek V4 Pro 0813 | DeepSeek 2026-08-12 | $0.111 / $0.097 | $4.44 | $3.9680 | |
| Grok 4.6 | xAI 2026-08-12 | $2 / $0.5 | $6 | $9.8000 |
How Prompt Caching & Batch Discounts Reshape 2026 LLM Economics
Across late-2026 frontier releases, standard uncached input rates and cached prefix read rates have diverged by up to 20x. Evaluating models on uncached sticker price alone leads to severe miscalculations for agentic and RAG workloads:
- OpenAI GPT-6.1 Terra Pro & Sol: Standard input is $1.00–$2.00 per 1M tokens, while cached input reads drop to $0.05–$0.10 per 1M tokens (a 95% reduction on repeated system prompts and tool definitions).
- Anthropic Claude Sonnet 5.5, Opus 5.5 & Haiku 5.5: Cached reads bill at 5% to 10% of base input cost ($0.10/1M on Sonnet 5.5; $0.01/1M on Haiku 5.5) across 1,000,000-token context windows.
- Google Gemini 4 Argon & 3.8 Flash: Context caching reduces 1M-token context reads from $2.00/1M down to $0.10/1M on Gemini 4 Argon.
- DeepSeek V4.1 Flash & GLM 5.3 Flash: Sub-cent cached input pricing ($0.018–$0.030/1M) enables high-throughput classification and extraction pipelines at a fraction of legacy model costs.
Exact Production Billing Formula
Single Request Cost = [ (Input_Tokens × (1 - Cache_Hit_Ratio) × Uncached_Input_Rate)
+ (Input_Tokens × Cache_Hit_Ratio × Cached_Input_Read_Rate)
+ (Output_Tokens × Output_Rate) ] ÷ 1,000,000 × Batch_Multiplier
30-Day Monthly Spend = Single Request Cost × Daily_Requests × 30
Related Engineering Utilities
Reduce your per-call token footprint by tightening tool definitions with our Strict JSON Schema & Zod Converter or configure local IDE connections with the MCP Server Config Builder.