Una hero card generata per GPT-6.1 Sol intitolata 'Near-Astra, una settimana dopo', con badge che riportano 'Rilasciato: 29 settembre 2026', 'Prezzo: $2,00 / $10,00 per 1M token' e 'Input in cache: $0,10, dimezzato', un footer che recita 'Dati OpenAI dichiarati dal fornitore; nessuna valutazione indipendente pubblicata al 30 settembre 2026', e il logo OrcaRouter nell'angolo in basso a destra.
Guides & Insights

GPT-6.1 Sol è disponibile: quasi-Astra a un quinto del prezzo, una settimana dopo il modello che sostituisce

Autore

Elias Hawthorne

Data di pubblicazione

Ultimi modelli · 20Vedi tutti i modelli →
Benchmark: Artificial Analysis · aggiornato ogni giorno
Torna a tutti gli articoli

Open​AI released GPT-6.1 Sol on September 29, 2026, at its DevDay 2026 keynote — a refresh of the GPT-6 Sol tier that shipped exactly one week earlier, repriced nowhere, and aimed squarely at a model costing five times as much. The headline numbers are unchanged from the model it replaces: $2.00 per million input tokens and $10.00 per million output tokens. The number that did move is the one most agent workloads actually pay: cached input fell from $0.20 to $0.10 per million tokens, half of GPT-6 Sol's rate and a tenth of GPT-6 Astra's. Open​AI's own framing is that GPT-6.1 Sol "nearly matches" GPT-6 Astra's intelligence on agentic coding, computer use and professional work at one-fifth of Astra's standard token prices, and that is a vendor claim, not a measured one — as of September 30 no independent evaluator has published a single score for the model. What follows separates the two.

Cosa è stato rilasciato e dove si trova il record

A screenshot of OpenAI's developer model page for GPT-6.1 Sol, showing the positioning line 'Near-Astra performance for complex work at a lower cost', a 1,050,000-token context window with 128,000 max output tokens and an Apr 30, 2026 knowledge cutoff, the pricing block reading $2.00 input, $0.10 cached input, $2.50 cache writes and $10.00 output per million tokens, and the note that reasoning.effort supports low, medium (default), high, xhigh and max while none and minimal are not supported.

The release is unusually well documented for a DevDay announcement. Within hours there was a model page at developers.openai.com, a pricing block, a system-card addendum at deploymentsafety.openai.com, and a commit in the vendor's own public API schema titled "Add gpt-6.1-sol model to model enums," timestamped 17:16 UTC on September 29. Open​AI's developer changelog, by contrast, still ends at the September 22 entry that announced GPT-6 Sol and GPT-6 Luna — the 6.1 entry has not been written up there, so the model page and the launch post are the primary record. The resolution matters because there is no second-generation snapshot to point at: the page's snapshot list contains one identifier, gpt-6.1-sol, with no dated variant, so "GPT-6.1 Sol" and the deployment you call today are the same thing.

La sua forma, fianco a fianco con ciò che sostituisce:

• Identificatore — gpt-6.1-sol, una singola snapshot rispetto a gpt-6-sol, anch'esso privo di variante datata
• Contesto — 1.050.000 token e 128.000 token massimi di output per entrambi; la pagina 6.1 indica esplicitamente entrambi i valori, mentre la pagina 6.0 documenta invece un massimo di 922.000 token di input
• Limite delle conoscenze — 30 aprile 2026 vs 20 aprile 2026
• Sforzo di ragionamento — basso, medio (predefinito), alto, xhigh, max, con none e minimal non supportati rispetto alla stessa scala più none
• Prezzi — $2,00 input / $0,10 in cache / $2,50 scrittura cache / $10,00 output per milione di token vs $2,00 / $0,20 / $2,50 / $10,00
• Prompt lunghi — oltre 272K token di input, il prezzo dell'intera richiesta viene ricalcolato con tariffe di input e cache a 2× e output a 1,5×, identico su entrambi i modelli
• Disponibilità — Plus, Pro, Business, Enterprise ed Edu in ChatGPT Work e Codex, oltre all'API; esplicitamente non nel prodotto Chat per consumatori
• Residenza — residenza dei dati negli Stati Uniti e nell'UE supportata, con modalità veloce non disponibile in caso di residenza nell'UE
• Strumenti — ricerca web, ricerca file, generazione immagini, interprete di codice, shell ospitata, apply patch, skill, computer use, MCP e ricerca strumenti, tutti supportati tramite l'API Responses
• Fine-tuning — non supportato su nessuno dei due

Una forma di API da segnalare prima di migrare: GPT-6.1 Sol supporta Chat Completions, ma il tool calling richiede la Responses API. Su GPT-6 Sol il vincolo era quello inverso-adiacente — il function calling in Chat Completions funzionava solo con reasoning_effort: "none". Se il tuo stack chiama gli strumenti tramite /v1/chat/completions, si tratta di una modifica al codice, non di un cambio di modello.

Le affermazioni del benchmark, etichettate come affermazioni

Every number in this section is Open​AI's, published in the launch post, and none of them has been reproduced by anyone outside Open​AI yet. They are worth reading in full because the pattern across them is consistent — the gains over GPT-6 Sol are real but the framing that travels is the Astra comparison, and the Astra comparison is always a cost comparison.

• DeepSWE v1.1, complex software-engineering tasks in real codebases — Open​AI reports GPT-6.1 Sol matching GPT-6 Astra at roughly one-fifth of the cost, and beating GPT-6 Sol's best score by 6.4 percentage points at a lower reasoning effort and lower cost.
• GDP.pdf, professional questions over complex PDF documents — Open​AI reports GPT-6.1 Sol scoring above Claude Opus 5.5 with fallbacks at less than half the cost per task across tested reasoning settings, and approaching Astra's state-of-the-art result at roughly one-fifth the cost per task.
• AutomationBench 1.0.6, multi-step business workflows across 47 tools — Open​AI reports 2.2 percentage points above Claude Opus 5.5 at medium reasoning effort at roughly a third of the cost, and 4.8 points above GPT-6 Sol at the same setting.
• OSWorld 2.0 offline set, long-horizon computer use — Open​AI reports a seven-point improvement over GPT-6 Sol at maximum reasoning effort at less than half the cost, and a result within 2.1 points of Astra at roughly one-seventh the cost per task.
• Terminal-Bench Science 0.1, scientific workflows — Open​AI reports GPT-6.1 Sol more than doubling GPT-6 Sol's score at maximum effort at less than half the cost per task, at $5.47 per task on average against $23.21 for Claude Opus 5.5 and $23.80 for Astra. Open​AI also states that Astra still holds the top score among models it tested at 68.1% and should be used for the hardest scientific work — a usefully un-marketing sentence.
• Factuality — on de-identified conversations where users had flagged an earlier model's error, Open​AI reports the share of responses containing a factual error falling from 11.4% to 7.7% at low reasoning effort, a reduction of about 32%, with the error rate staying within 1.9 points of Astra's across tested settings at less than one-fifth the cost per task.

Two of those deserve translation. First, the factuality evaluation runs on prompts deliberately chosen because a previous model got them wrong; Open​AI says so in the post, and it means the 11.4% and 7.7% describe a hostile slice of traffic, not your traffic. Second, the paragraph that reads most strongly for the model is the one about search-tool transparency in the safety addendum: on tasks built to elicit failures at maximum reasoning effort, GPT-6.1 Sol fails to tell the user that its search tool is broken in 2.1% of cases, against 4.9% for GPT-6 Sol, 1.5% for GPT-6 Astra and 28.7% for GPT-6 Luna. That is the kind of number that decides whether an agent is deployable, and it is also, again, the vendor's own.

Il registro indipendente è vuoto, e questa è l'altra metà della storia.

A screenshot of the OrcaRouter model catalogue filtered to the query gpt-6, headed '3 models - 1 providers - one API key, one bill', showing three OpenAI cards side by side: openai/gpt-6-luna described as the fast, cost-efficient model in the GPT-6 series, openai/gpt-6-sol described as the cost-efficient high-end model positioned below the flagship, and openai/gpt-6-astra described as the flagship for demanding long-horizon work. Each card carries vision, tools and reasoning tags and a base-rate block; the GPT-6 Sol card reads $2.00 input, $10.00 output, $0.200 cache read and $2.50 cache write per million tokens.

Artificial Analysis — la classifica di terze parti che questo blog considera il riferimento neutrale — non ha alcuna voce per GPT-6.1 Sol. I suoi URL dei modelli per lo slug 6.1 Sol restituiscono 404, e la stringa non compare nell'HTML della classifica live. Ciò che invece la classifica ha è una valutazione completa di GPT-6 Sol al massimo sforzo di ragionamento: un Intelligence Index di 48, un costo di 1,06 $ per attività dell'indice, 77 milioni di token di output generati durante l'esecuzione dell'indice rispetto a una mediana della classifica di 88 milioni, e un Coding Agent Index di 57 a 2,99 $ per attività. Quelle cifre sono l'ancora indipendente più vicina per il modello che GPT-6.1 Sol sostituisce, e sono ciò rispetto a cui qualsiasi affermazione su 6.1 dovrebbe prima o poi essere misurata.

Until that entry appears, anyone quoting a 6.1 Sol score against another model is quoting Open​AI against Open​AI. That is not a reason to distrust the launch post; it is a reason to run your own evaluation set before you move a production path. The vendor gives an unusually concrete instruction on this itself: Open​AI tells developers already using GPT-6 Sol to review the migration guidance before switching, and the migration guidance tells them to compare configurations on representative tasks rather than assuming the highest effort is the best trade-off.

Il taglio sull'input in cache è la parte che cambia davvero la fattura.

Un listino prezzi che non si muove ma una tariffa di cache che si dimezza è facile da trascurare, ed è la riga di maggior impatto del rilascio per chiunque gestisca agenti. L'input in cache a 0,10 $ per milione di token è il 5% della tariffa dell'input non in cache, la metà della tariffa in cache di GPT-6 Sol, e — rispetto all'input in cache di GPT-6 Astra a 1,00 $ — un decimo di quella del modello di punta. I loop degli agenti che reinviano un prefisso stabile migliaia di volte vivono quasi interamente su quella riga. Lo stesso mese-agente da 200 milioni di token che su GPT-6 Sol verrebbe fatturato a circa 40 $ di input in cache, su GPT-6.1 Sol ne costa circa 20 $; su GPT-6 Astra lo stesso traffico in cache costa 200 $.

The arithmetic never gets that clean in production, because cache reads only bill at that rate when the prefix actually hits, and Open​AI notes that changing reasoning effort or the available tool set no longer invalidates a cache entry the way it once did — which is the quiet fix that makes the cache line usable in an agent loop at all. The other threshold to watch is unchanged: cross 272,000 input tokens and the entire request reprices at twice the input and cache rates and 1.5× the output rate. Batch and Flex run 50% below standard, fast mode runs at 2× standard, and regional processing adds a 10% premium where it is available.

Cosa non è ancora stato rilasciato e cosa significa per il routing

A generated scoreboard titled 'GPT-6.1 Sol - the scoreboard' listing six rows: price $2.00 input and $10.00 output per million tokens; cached input $0.10 per million, halved from GPT-6 Sol; context window 1,050,000 tokens; independent score none published yet; availability in ChatGPT Work, Codex and the API; vendor claim near-Astra at about one fifth the cost. A footer reads that OpenAI figures are vendor-reported and no independent index score was published as of September 30, 2026.

Open​AI says GPT-6.1 Sol Ultrafast arrives "in the coming days" in Codex at up to 8× faster token generation, with no price attached. That is a promise, not a product, and it is worth naming as such: there is no Ultrafast rate card for the 6.1 tier today, and Open​AI has not said whether the 8× figure is measured against standard speed or against fast mode, which is itself priced at 2× standard. Treat the announcement as a scheduling signal and price it when a number appears.

On our side, honestly: GPT-6.1 Sol is not on OrcaRouter's routes yet. The public catalogue endpoint returns "model not found" for openai/gpt-6.1-sol today, and our Open​AI slate carries GPT-6 Astra, GPT-6 Sol and GPT-6 Luna. What is callable on OrcaRouter right now is GPT-6 Sol at Open​AI's own list price with zero markup — the $2.00 / $0.20 / $10.00 rate card, and the 272K repricing rule, passed through exactly as the vendor lists it. That matters more than usual this week for a specific reason: the same pass-through means if Open​AI cuts the 6.1 tier's price, or attaches a number to Ultrafast, the change is live on our side the same day it is live on Open​AI's, with no second contract to renegotiate. When 6.1 Sol lands on the routes it will appear in the catalogue at the vendor's price like everything else, and until then the honest answer is that the model it replaces is the one you can call.

La versione pratica della decisione: se sei già su GPT-6 Sol e il tuo carico di lavoro è pesante sulla cache, agentico o a lungo orizzonte, il refresh 6.1 migliora esattamente le dimensioni per cui stai pagando, e la migrazione è un cambio di stringa del modello più un passaggio da Chat-Completions a Responses se usi strumenti. Se il tuo carico di lavoro è generazione con prompt brevi e poco riutilizzo, il rilascio non cambia quasi nulla per te — stesso prezzo di input, stesso prezzo di output, stessa finestra. E se stavi aspettando un punteggio di terze parti prima di impegnarti, quell'attesa è ancora aperta.

Confrontati in questo articolo2

Rilevato da questo articolo · Benchmark: Artificial Analysis · aggiornato ogni giorno