OpenRouter · free and discounted models · live pricing
Free & discounted OpenRouter models
Fetched directly from the public OpenRouter model catalog at 2026-09-06T14:43:07.379Z. Free means the prompt and completion prices are both exactly zero — read from pricing, not from the :free suffix, which is a naming habit rather than the fact. Unknown pricing is never treated as free. Discountis the conditional pricing OpenRouter publishes per model (off-peak UTC windows and volume tiers), shown as the saving against that model’s most expensive rate; the Input column is the headline rate, which for these models is the cheap end of the range. Separately, "Cheapest 20%" marks models whose combined prompt+completion price is at or below $0.62 / 1M tokens — a ranking, not a discount.
Published discount is a different fact again: a straight percentage off list price that a named provider publishes for its endpoint. It is not on the model — no model in the catalog carries a discount key — it is one level down, on the provider endpoint, which is why this page previously reported that OpenRouter published no discounts at all. Endpoints are collected under a daily request budget, so coverage is partial: 52 models have an observed endpoint and 4 of those carry a non-zero discount. A dash means no endpoint has been observed for that model yet — it does not mean there is no discount.
+48 more providers
Models (430)
Price per million tokens, sorted by the selected criteria. 50 per page.
| Model | Provider | Input $/1M | Output $/1M | Context | Discount | Published discount | Status | Vision |
|---|---|---|---|---|---|---|---|---|
| Qwen: Qwen3 8B | qwen | $0.12 | $0.45 | 131,072 | — | — | Cheapest 20% | No |
| Qwen: Qwen3.8 Flash | qwen | $0.15 | $0.47 | 1,000,000 | — | — | Cheapest 20% | Yes |
| Qwen: Qwen3 30B A3B | qwen | $0.12 | $0.50 | 131,072 | — | — | Cheapest 20% | No |
| Qwen: Qwen3 235B A22B Instruct 2507 | qwen | $0.09 | $0.55 | 262,144 | — | — | — | No |
| Z.ai: GLM 5.3 Flash (batch) | z-ai | $0.15 | $0.50 | 1,048,575 | — | — | — | Yes |
| Tencent: Hy3 | tencent | $0.13 | $0.53 | 262,144 | −37%00:00–16:00 UTC, 16:00–00:00 UTC | — | — | No |
| DeepSeek: DeepSeek V3.2 | deepseek | $0.27 | $0.40 | 163,840 | — | — | — | No |
| DeepSeek: DeepSeek V3.2 Exp | deepseek | $0.27 | $0.41 | 163,840 | — | — | — | No |
| OpenAI: GPT-5.6 Luna Pro (batch) | openai | $0.10 | $0.60 | 1,050,000 | −50%≥ 272K prompt tokens | — | — | Yes |
| OpenAI: GPT-5.6 Luna (batch) | openai | $0.10 | $0.60 | 1,050,000 | −50%≥ 272K prompt tokens | — | — | Yes |
| Tencent: Hunyuan A13B Instruct | tencent | $0.14 | $0.57 | 131,072 | — | — | — | No |
| OpenAI: GPT-5.4 Nano (batch) | openai | $0.10 | $0.63 | 400,000 | — | — | — | Yes |
| Mistral: Mistral Small 4 | mistralai | $0.15 | $0.60 | 262,144 | — | — | — | Yes |
| Upstage: Solar Pro 3 | upstage | $0.15 | $0.60 | 131,072 | — | — | — | No |
| Qwen: Qwen3 VL 30B A3B Instruct | qwen | $0.15 | $0.60 | 262,144 | — | — | — | Yes |
| OpenAI: gpt-oss-120b (batch) | openai | $0.15 | $0.60 | 131,072 | — | — | — | No |
| Cohere: Command R (08-2024) | cohere | $0.15 | $0.60 | 128,000 | — | — | — | No |
| OpenAI: GPT-4o-mini | openai | $0.15 | $0.60 | 128,000 | — | — | — | Yes |
| OpenAI: GPT-4o-mini (2024-07-18) | openai | $0.15 | $0.60 | 128,000 | — | — | — | Yes |
| Qwen2.5 72B Instruct | qwen | $0.36 | $0.40 | 32,768 | — | — | — | No |
| Tencent: Hy3 preview | tencent | $0.18 | $0.60 | 262,144 | — | — | — | No |
| Mistral: Saba | mistralai | $0.20 | $0.60 | 32,768 | — | — | — | No |
| TheDrummer: UnslopNemo 12B | thedrummer | $0.40 | $0.40 | 1,024,000 | — | — | — | No |
| Meta: Llama 3.1 70B Instruct | meta-llama | $0.40 | $0.40 | 131,072 | — | — | — | No |
| TheDrummer: Cydonia 24B V4.1 | thedrummer | $0.30 | $0.50 | 131,072 | — | — | — | No |
| Google: Gemini 3.1 Flash Lite (batch) | $0.13 | $0.75 | 1,048,576 | — | — | — | Yes | |
| DeepSeek: DeepSeek V4 Flash Vision Exp | deepseek | $0.22 | $0.66 | 1,048,576 | −50%00:00–01:00 UTC, 04:00–06:00 UTC, 10:00–00:00 UTC | — | — | Yes |
| Meta: Llama 4 Maverick | meta-llama | $0.20 | $0.70 | 1,048,576 | — | — | — | Yes |
| Mistral: Mistral Small 3.1 24B | mistralai | $0.35 | $0.55 | 128,000 | — | — | — | Yes |
| Qwen: Qwen3 Coder Next | qwen | $0.12 | $0.80 | 262,144 | — | — | — | No |
| Z.ai: GLM 4.5 Air | z-ai | $0.13 | $0.85 | 131,072 | — | — | — | No |
| Qwen: Qwen3.6 35B A3B | qwen | $0.10 | $0.90 | 262,144 | — | — | — | Yes |
| OpenAI: GPT-4.1 Mini (batch) | openai | $0.20 | $0.80 | 1,047,576 | — | — | — | Yes |
| Inception: Mercury 2 | inception | $0.25 | $0.75 | 128,000 | — | — | — | No |
| ReMM SLERP 13B | undi95 | $0.35 | $0.65 | 6,144 | — | — | — | No |
| OpenAI: GPT-3.5 Turbo (batch) | openai | $0.25 | $0.75 | 16,385 | — | — | — | No |
| Qwen: Qwen Plus 0728 | qwen | $0.26 | $0.78 | 1,000,000 | −67%≥ 256K prompt tokens | — | — | No |
| Qwen: Qwen-Plus | qwen | $0.26 | $0.78 | 1,000,000 | −67%≥ 256K prompt tokens | — | — | No |
| Arcee AI: Trinity Large Thinking | arcee-ai | $0.25 | $0.80 | 262,144 | — | — | — | No |
| Venice: Uncensored | cognitivecomputations | $0.20 | $0.90 | 128,000 | — | — | — | No |
| OpenAI: GPT-5 Mini (batch) | openai | $0.13 | $1.00 | 400,000 | — | — | — | Yes |
| Mancer: Weaver (alpha) | mancer | $0.40 | $0.75 | 8,000 | — | — | — | No |
| Qwen: Qwen3 Coder Flash | qwen | $0.20 | $0.97 | 1,000,000 | −62%≥ 32K prompt tokens, ≥ 128K prompt tokens | — | — | No |
| Z.ai: GLM 4.6V | z-ai | $0.30 | $0.90 | 131,072 | — | — | — | Yes |
| Mistral: Codestral 2508 | mistralai | $0.30 | $0.90 | 256,000 | — | — | — | No |
| Qwen: Qwen3 Next 80B A3B Instruct | qwen | $0.10 | $1.10 | 262,144 | — | — | — | No |
| DeepSeek: DeepSeek V3 | deepseek | $0.32 | $0.89 | 163,840 | — | — | — | No |
| WizardLM-2 8x22B | microsoft | $0.62 | $0.62 | 65,535 | — | — | — | No |
| Nex AGI: Nex-N2-Pro | nex-agi | $0.25 | $1.00 | 262,144 | — | — | — | Yes |
| DeepSeek: DeepSeek V3 0324 | deepseek | $0.25 | $1.00 | 163,840 | — | — | — | No |