Pricing
9 articles
-
OpenAIGPT-6.1 Terra Obeyed Peer Agents in OpenAI's Tests, Not Mine
OpenAI's GPT-6.1 Terra system card says the model performed an unauthorized action requested by an apparent peer agent in 11 of 80 samples, against 1 for GPT-6.1 Sol. Across 780 runs of a three-scenario peer-board test (420 Terra, 360 Sol; blunt asks, no written policy, relayed human approval, and requests for a promise), neither model did it once or promised to. At max effort Terra replied to the peer in 54 of 60 runs against 10 of 60 for Sol, and its extra turns and reasoning made it cost as much per run as Sol ($0.0077 and $0.0086 vs $0.0074 and $0.0071); at medium it was half the price as advertised.
-
OpenAIOpenAI's 12-Hour Cache Can Make You Pay Twice
On 20 October OpenAI added a 12-hour minimum prompt-cache lifetime to the Responses API at 2.5x the input price per write, against 1.25x for the default 30 minutes, with reads priced the same under both. Every read extends a 30-minute entry for free, so the longer lifetime pays off only when the prefix would have gone cold at least twice in 12 hours. On a 7,131-token prefix on gpt-6-luna, re-requesting a 30-minute-cached prefix six seconds later with ttl 12h billed the full 12-hour write with zero cached tokens, 3.75x in total, as the guide documents. The Responses reference still calls 30m the only supported value, a bad value's error lists only 30m, Chat Completions accepts the option but rejects 12h and the per-block ttl its own reference documents, prewarm writes the 12-hour entry with no output, and LiteLLM prices 12-hour writes at the 30-minute rate until PR #47800 merges.
-
OpenAIOpenAI Retires GPT-4o Early. Its Replacement Rejects Temperature.
On 14 October OpenAI told developers that GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini leave the API on 14 January 2027, 92 days later, while the same deprecations page promises generally available models at least six months of notice and gives no reason for the shorter window. The replacement named for all four is gpt-6-luna at $0.10 and $0.50 per million tokens, a reasoning model that defaults to medium effort: swapping the model ID into a GPT-4o-style Chat Completions call returns a 400 for temperature, a 400 for a function tool and a 400 for max_tokens until reasoning_effort is set to none, and at its default it costs more per call than GPT-4o mini on a one-word classifier. Fine-tuned versions die the same day, new fine-tuning jobs end on 6 January, Luna cannot be fine-tuned, Predicted Outputs is left with no model, and Azure runs the same snapshots until 14 April 2027.
-
MetaMuse Wonder Full Isn't the Model You Get on the API
Meta released Muse Wonder and Muse Wonder Full on 12 October. Artificial Analysis scored Full at 56 on its Intelligence Index, two points behind Claude Opus 5.5, and Wonder at 52, but only Wonder is in the Meta Model API, at $4 per million input and $20 per million output, the same as Opus 5.5. Full is 'Not available' on the API and sold only on the $50 and $200 Muse Code plans, and Meta's models page says that setting reasoning effort to max on Wonder doesn't select it. At the Opus price Wonder costs $4.52 per index task, 76% of it input from 7.69 million input tokens per task, and the subscription that carries Full defaults to Contributor mode, where your prompts, code and files train future Meta models.
-
GoogleGemini 4 Argon Is Cheap Per Token, Not Per Task
Google made Gemini 4 Argon generally available on 13 October at an introductory $2 per million input tokens and $10 per million output, a fifth of GPT-6 Astra and half of Claude Opus 5.5, with the price doubling to $4 and $20 on 1 December. Artificial Analysis scored it level with Astra on its Intelligence Index but found it costs 60% of Astra per task, not 20%, because Argon averages 62,000 output tokens per task against Astra's 27,000; at the standard price that becomes $3.98 per task against Astra's $3.26. The migration guide also makes max_output_tokens a hard cap on thinking plus output and tells you to raise it to at least 32,768, silently ignores temperature, top_p and top_k, and drops explicit caching entirely.
-
OpenAIOpenAI's Decisions API Is Fast and Overconfident
OpenAI opened the Decisions API to all developers on 6 October: POST /v1/decisions on gpt-6-luna returns a probability, a choice or a score 'about 10x faster than the Responses API' at $0.10 per million input tokens with no output or cache charges. A forum tester ran a 70%-heads coin through it: the predicate returned 70% over 1,000 trials, the choice format returned heads 98% of the time and moved when options were reordered. The guide never defines its confidence field or mentions calibration. With no caching, a classifier with a 4,000-token rubric costs about five times more here than Luna through Responses; the endpoint only wins when output exceeds 18% of the cached prefix.
-
AnthropicHaiku 5.5 Is 10x Cheaper Until Its Tokenizer Catches Up
Claude Haiku 5.5 launched on 7 October at $0.10/$0.50 per million tokens, a tenth of Haiku 4.5, but past 100,000 prompt tokens every rate is five times higher and applies to the whole request, cached tokens included. The newer tokenizer counts about 30% more tokens, so a 77,000-token Haiku 4.5 prompt lands over the line unchanged. It is the only Claude model priced by length; OpenAI's line is at 272K at 2x, Gemini 3.8 Flash has none. Gateways mispriced it, Haiku 4.5's retirement floor ends 15 October, and temperature, prefill and budget_tokens now return 400.
-
AgentsMeta Cut Coding Agent Prices 12x. You Pay in Data.
Meta shipped Muse Code on 5 August with the same model behind two IDs: muse-spark-1.2 at $1.25 per million input tokens, and muse-spark-1.2-contributor at $0.10, where your traffic may be used to train Meta's models. The discount is not a smaller model, it is a licensing decision made by editing one string. And in a harness with a 1M-token window, the prompt is whatever the agent decided to read.
-
InferenceGPT-5.6 Rewrote Its Own GPU Kernels
OpenAI cut GPT-5.6 Luna 80 percent on July 30, three weeks after general availability, funded by Sol rewriting OpenAI's own production GPU kernels and its speculative-decoding draft model. The transferable part is not the discount, it is FpSan, the floating-point sanitizer built to verify kernels no human read line by line.