AI MODEL ATLAS · DECISION PAGE
Fastest LLM API models by OpenRouter order
OpenRouter models in the server-provided throughput-high-to-low order.
OBJECT ID
asca:openrouter-model-page-fastest-llm-api@2026-09-01.v1FETCHED
STATUSREPRODUCIBLE SNAPSHOT
QUESTION · METHOD
Fastest LLM API models by OpenRouter order
Preserve server-provided order; do not interpret rank as a latency or tokens-per-second value.
OBSERVED MODELS · 10 SHOWN
Comparable fields from one dated catalog snapshot
| Rank / order | Model | Input / 1M | Output / 1M | Request | Context | Tools | Structured |
|---|---|---|---|---|---|---|---|
| 1 | Google: Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) | $0.25 | $1.50 | $0.00 | 65.54K | No | No |
| 2 | OpenAI: gpt-oss-safeguard-20b | $0.075 | $0.30 | $0.00 | 131.07K | Yes | Yes |
| 3 | Morph: Morph V3 Fast | $0.80 | $1.20 | $0.00 | 81.92K | No | No |
| 4 | Inception: Mercury 2 | $0.25 | $0.75 | $0.00 | 128K | Yes | Yes |
| 5 | SpaceXAI: Grok 4.20 Multi-Agent | $1.25 | $2.50 | $0.00 | 2M | No | Yes |
| 6 | LiquidAI: LFM2.5-2.6B (free) | $0.00 | $0.00 | $0.00 | 65.54K | Yes | Yes |
| 7 | Qwen: Qwen3.5-35B-A3B | $0.25 | $1.25 | $0.00 | 262.14K | Yes | Yes |
| 8 | Qwen: Qwen3 Next 80B A3B Thinking | $0.15 | $1.20 | $0.00 | 262.14K | Yes | Yes |
| 9 | Meta: Llama 4 Scout | $0.11 | $0.34 | $0.00 | 1.31M | Yes | Yes |
| 10 | Poolside: Laguna XS 2.1 | $0.06 | $0.12 | $0.00 | 262.14K | Yes | No |
Prices are listed USD per token converted to USD per 1M tokens. Request fee is a separate listed field; it is not included in derived text-token cost. A free listing means zero listed token price in this snapshot, not unlimited access.
RANGE · BIAS · LIMIT · CONFIDENCE
- Observed range: 425 matching catalog records; 10 shown.
- Bias: OpenRouter measurement and catalog bias.
- Limitation: No numeric throughput, latency, hardware, prompt size, or reproducibility claim is made.
- Confidence: medium for the recorded response and stated transformation.
DOWNLOAD · VERIFY · CITE
Keep the snapshot and object ID with the decision.
“Preserve server-provided order; do not interpret rank as a latency or tokens-per-second value.”— asca:openrouter-model-page-fastest-llm-api@2026-09-01.v1
RELATED DECISION PAGES
- Cheapest paid LLM API models
- Long-context LLMs for a 128K + 4K workload
- Tool-calling compatible models
- Structured-output compatible models
- Vision-language models
- Free LLM API models
- Most popular AI models by OpenRouter weekly order
- Fastest LLM API models by OpenRouter order
- Coding AI models by OpenRouter order
- Agentic AI models by OpenRouter order
- Lowest-cost models among OpenRouter's top 100 coding order
- Lowest-cost models among OpenRouter's top 100 agentic order
- Tool-calling models with at least one million context tokens
- Image-input models with tool-calling support
- Models declaring both tools and structured outputs