In CC Switch Routing mode, Codex gets an empty model list and shows its bundled one, which has GPT-6.1 Terra only from 0.165.0. Go back to Direct and restart.
Google announced Gemini 4 Carbon on October 28, 2026, but only Fairwind cyber defenders can use it. No public API or app date yet; build on Argon meanwhile.
GPT Image 2 costs about $0.005 to $0.43 per image in output on OpenAI's API, set by size and quality. Batch halves it, and ChatGPT has no per-image price.
As of October 29, 2026, $0.03 per image on LaoZhang API is the lowest price found for native 4K GPT Image 2 and 2.5, at every quality tier, edits included.
Claude Haiku 5.5 and GPT-6 Luna share $0.10/$0.50 rates under 100K tokens, yet Haiku cost 2.2x more per task at default effort and reprices 5x above 100K.
Claude Haiku 5.5 costs 1/20 of Sonnet 5.5 per input and output token but trails it on agentic coding. Move checkable work to Haiku; keep agents on Sonnet.
GPT-6.1 Terra's 1/9 cost claim is max vs max. At equal AA scores, Opus 5.5 high costs 3.6x Terra max per task, and English input bills about 5.8x.
Against GPT-6.1 Sol, Opus 5.5 costs exactly 2x per token under 272K input and stays pricier above. Opus scores higher; pay only where it passes more tasks.
Sonnet 5.5 keeps Sonnet 5's $2/$10 token prices. In Artificial Analysis's October 5 tests it cost less per task from low to xhigh and about 51% more at max.
Same $2/$10 list price, but Sonnet 5.5 costs 2.2–7.9x GPT-6.1 Sol per AA task. Start on Sol; use Sonnet for terminal agents and prompts over 272K tokens.
GPT-6 Astra outscores Sonnet 5.5 at every effort level on Artificial Analysis' index. Sonnet is cheaper per task up to high, faster, and wins above 272K input.
Sonnet 5.5 costs half as much per token and suits well-scoped work at low or medium effort. At max it costs as much per task as Opus 5.5 and scores lower.
Gemini 4 Argon: resend with continuationToken on CONTINUATION, raise the budget when MAX_TOKENS has no text. OpenAI-compatible endpoints can't continue.
Gemini 4 Argon opened to paid API users on October 13 at $2/$10, doubling December 1. At matched effort it costs more per task than Opus 5.5 and 6.5x GPT-6.1 Sol.
Gemini 4 Argon lists at 2.67x 3.8 Flash per token, but on five same-prompt calls it cost 1.57x as much through Nov 30, 2026, and 3.15x in December.
As of October 14, 2026, you can call GPT-6.1 Sol and Gemini 4 Argon, available since October 13. Fable 5.5 is unannounced, and GPT-6.1 Astra was shelved.
Both list at $10/$50 per million tokens, but on October 26, 2026 Fable 5.1 counted 1.45x Astra's tokens for English and cost 1.5–1.6x on tasks both passed.
Move gpt-6-sol traffic to GPT-6.1 Sol at the same price. Astra leads it by 2–8 points at equal effort; Sol max ties Astra high for 2.6x less per task.
Terra's API and Codex rates are half of Sol's, and it ran small Codex tasks twice as fast. At xhigh effort, though, it costs more per task than Sol at medium.
GPT-6.1 Terra costs 10× GPT-6 Luna per token. On 12 tasks tested October 28, 2026, it cost 2.7–6.3× as much per passed task and passed 15–16 of 16 to Luna's 12.
GPT-6 now has a Terra: GPT-6.1 Terra, released October 26, 2026, costs half of GPT-6.1 Sol per token and about 10 times Luna. Pick by workload, then test.
K3.1 Pro's $2/$10 is half of Opus 5.5's $4/$20 per token, yet at Kimi's default max effort a task can cost as much as Opus. Set effort before you switch.
Both list at $2/$10 per 1M tokens through November 30, 2026, then Argon doubles. K3.1 Pro leads terminal coding; Argon leads overall and on factual recall.
Both list at $2/$10 per million tokens, but Artificial Analysis puts K3.1 Pro at $3.68 per task against $1.27 for Sol at max, which also scores higher.