spb/llm-api-toolkits

1<!-- Generated by build-skill.mjs from live API probes and official documentation (2026-11-10). Do not edit by hand. -->2# Models34Ids and behaviour observed 2026-11-10. Context window, output cap: official model pages. Prices: official pricing page, USD per million tokens (standard tier, prompts up to 272K tokens).56Recommended: **gpt-6-astra** (flagship), **gpt-6.1-sol** (balanced default), **gpt-6.1-terra** (fast), **gpt-6-luna** (cheapest).78| Model | Context | Max output | Input | Cached input | Output | Batch in / out |9|---|---|---|---|---|---|---|10| `gpt-6-sol` | 1.05M | 128K | $2 | $0.2 | $10 | $1 / $5 |11| `gpt-6.1-terra` | 1.05M | 128K | $1 | $0.05 | $5 | $0.5 / $2.5 |12| `gpt-6.1-sol` | 1.05M | 128K | $2 | $0.1 | $10 | $1 / $5 |13| `gpt-6-luna` | 1.05M | 128K | $0.1 | $0.01 | $0.5 | $0.05 / $0.25 |14| `gpt-5.6-terra` | 1.05M | 128K | $2 | $0.2 | $12 | $1 / $6 |15| `gpt-5.6-sol` | 1.05M | 128K | $4 | $0.4 | $20 | $2 / $10 |16| `gpt-6-astra` | 1.05M | 128K | $10 | $1 | $50 | $5 / $25 |17| `gpt-5.5` | 1.05M | 128K | $5 | $0.5 | $30 | $2.5 / $15 |18| `gpt-5.4` | 1.05M | 128K | $2.5 | $0.25 | $15 | $1.25 / $7.5 |19| `gpt-5.4-mini` | 400K | 128K | $0.75 | $0.075 | $4.5 | $0.375 / $2.25 |20| `gpt-5.4-nano` | 400K | 128K | $0.2 | $0.02 | $1.25 | $0.1 / $0.625 |21| `gpt-5.3-codex` | 400K | 128K | — | — | — | — |22| `gpt-5.2` | 400K | 128K | $1.75 | $0.175 | $14 | $0.875 / $7 |23| `gpt-5.6-luna` | 1.05M | 128K | $0.2 | $0.02 | $1.2 | $0.1 / $0.6 |24| `gpt-5.1` | 400K | 128K | $1.25 | $0.125 | $10 | $0.625 / $5 |25| `gpt-5.2-pro` | 400K | 128K | $21 | — | $168 | $10.5 / $84 |26| `gpt-5-mini` | 400K | 128K | $0.25 | $0.025 | $2 | $0.125 / $1 |27| `gpt-5-nano` | 400K | 128K | $0.05 | $0.005 | $0.4 | $0.025 / $0.2 |28| `gpt-5` | 400K | 128K | $1.25 | $0.125 | $10 | $0.625 / $5 |29| `gpt-4.1` | 1.05M | 33K | $2 | $0.5 | $8 | $1 / $4 |30| `gpt-4.1-mini` | 1.05M | 33K | $0.4 | $0.1 | $1.6 | $0.2 / $0.8 |31| `gpt-4o` | 128K | 16K | $2.5 | $1.25 | $10 | $1.25 / $5 |32| `gpt-4o-mini` | 128K | 16K | $0.15 | $0.075 | $0.6 | $0.075 / $0.3 |33| `o3` | 200K | 100K | $2 | $0.5 | $8 | $1 / $4 |34| `gpt-5-pro` | 400K | 272K | $15 | — | $120 | $7.5 / $60 |35| `chat-latest` | 400K | 128K | — | — | — | — |36| `gpt-5.4-pro` | 1.05M | 128K | $30 | — | $180 | $15 / $90 |37| `gpt-5.5-pro` | 1.05M | 128K | $30 | — | $180 | $15 / $90 |38| `o3-pro` | 200K | 100K | $20 | — | $80 | $10 / $40 |3940Notes:41- "—" means the official pricing page lists no price; the scripts report it as unknown rather than guess.42- Above 272K input tokens the newest models switch to long-context rates (about 2× input, 1.5× output). `cost` applies them.43- Fast mode (formerly priority processing) costs about 2×; flex and batch about half.44- Reasoning tokens are billed as output tokens and count against `max_output_tokens`.45- Embeddings: `text-embedding-3-small` $0.02, `text-embedding-3-large` $0.13 per million tokens. Moderation is free.46- Dated snapshots (`gpt-5.4-2026-03-05`) resolve to their alias in the scripts.47- The models endpoint returns ids only. Never write an id from memory: run `node scripts/openai-api.mjs models`.48