Prices the docs attach to coding models
This page does not pick a cheap coding model. It lists prices and context labels from the provider pages linked on the cost calculator, and it repeats a coding or tool sentence only when that page says one. “Cheap”, “mini”, and “flash” are used only when the provider used them.
| Id | Standard input / output per 1M | Context as printed | What the page says about coding |
|---|---|---|---|
| gpt-6-luna | $0.10 / $0.50 | 1,050,000 tokens | Called cost-sensitive and efficient. The model page does not call it a coding model. Function calling is documented, and Chat Completions function calling requires reasoning_effort none. |
| gpt-6.1-sol | $2 / $10 | 1,050,000 tokens | The models index says it balances intelligence and cost. The model page fetched here does not add a coding slogan. Long prompts over 272K input tokens use 2× input and 1.5× output for the full request. |
| claude-haiku-5-5 | $0.10 / $0.50 up to 100,000 prompt tokens; $0.50 / $2.50 over that | 1M tokens | Overview: high-volume, latency-sensitive tasks such as classification, extraction, and routing. Not described there as a coding model. |
| claude-opus-5-5 | $4 / $20 | 1M tokens | Overview: long-running agentic coding and knowledge work. |
| openai/gpt-oss-20b on Groq | $0.075 / $0.30 | 131,072 tokens | The production-models row does not contain a coding sentence. |
| openai/gpt-oss-120b on Groq | $0.15 / $0.60 | 131,072 tokens | Featured text: built-in browser search and code execution, and reasoning. |
| gpt-oss-120b on Cerebras | $0.35 / $0.75 | 65k free tier, 131k paid | Model card: efficient reasoning across science, math, and coding, and real-time coding assistance. |
| mistral-small-2603 | $0.15 / $0.60 | 256k | Card: instruct, reasoning, and coding in one model. function-calling is a listed feature. |
| codestral-2508 | $0.30 / $0.90 | 128k | Card: fill-in-the-middle and code generation. |
| zai-org/GLM-5.3-Flash on Together | $0.15 / $0.50 | 1M | Page heading: Frontier Coding at Flash Cost. Cached input is $0.03 / 1M. |
The calculator uses the same standard rates and ignores cache, batch, and long-context multipliers. Those multipliers are in each row’s note on the calculator page. OpenRouter does not publish a single hosted price for these ids; its FAQ says inference is passed through at the provider rate, with a separate fee when you buy credits.
Other notes
- Use OpenRouter with Cline
- Use Ollama local models with OpenCode
- Use Groq with Cline and OpenCode
- Token price calculator
- LLM APIs category list
- Free LLM API limits
Sources
Last checked 2026-10-22.