Comparison · verified November 2, 2026
Claude vs GPT-6 for Customer Support
Support is the one application where the cheapest capable model usually wins, because volume dominates everything else. It is also the application where a confidently wrong answer costs the most, because it reaches a customer with your name on it.
Those two facts pull in opposite directions, and that tension — not benchmark scores — is what should decide this. Below: verified pricing at real ticket volume, the behavioural differences that matter for escalation, and a shadow-mode test to run before either model touches a live conversation.
The verified differences
What each one costs
Last verified November 2, 2026
USD per million tokens, standard rates, no discounts applied. Read the three caveats underneath before you put these numbers in a spreadsheet.
| Model | Tier | Input | Output | Notes |
|---|---|---|---|---|
| Claude Fable 5.1 | frontier | $10 | $50 | The frontier model, for long-running agents and the hardest reasoning. Cache hits are 2.5% of input. |
| Claude Opus 5.5 | frontier | $4 | $20 | The default flagship. Cheaper than Opus 5 ($5/$25). Fast mode is $8/$40. |
| Claude Sonnet 5.5 | balanced | $2 | $10 | What most production agent fleets run on. Replaces Sonnet 4.5, which retires November 30, 2026. |
| Claude Haiku 5.5 | fast | $0.10 | $0.50 | High-volume and latency-sensitive work. Prompts over 100K tokens meter at $0.50/$2.50. |
| GPT-6 Astra | frontier | $10 | $50 | OpenAI’s flagship since September 2026. Requests over 272K tokens are billed at a higher rate for the whole request. |
| GPT-6.1 Sol | balanced | $2 | $10 | The everyday tier, shipped at DevDay on September 29, 2026. |
| GPT-6.1 Terra | balanced | $1 | $5 | A cheaper mid tier. |
| GPT-6 Luna | fast | $0.10 | $0.50 | The low-cost tier. Same list price as Claude Haiku 5.5. |
Anthropic changed tokenizers at Claude 4.7
Claude 4.7 and later, including every 5.x model, use a newer tokenizer that produces roughly 30% more tokens for the same text than Sonnet 4.6 and earlier. A $2/MTok model on the new tokenizer is not directly comparable to a $2/MTok model on an older one, or to another vendor. Compare cost per task, not cost per token.
Long context costs extra on some models
GPT-6 Astra bills the whole request at a higher rate once it passes 272K tokens, Gemini 3.1 Pro charges more for prompts over 200K tokens, and Claude Haiku 5.5 charges more for prompts over 100K tokens. Claude Fable, Opus and Sonnet include the full 1M-token window at standard pricing, so a 900k-token request costs the same per token as a 9k one. If your workload is context-heavy, check the long-context rate before the headline rate.
Google’s current rates are introductory
Gemini 4 Argon doubles to $4/$20 on December 1, 2026, and Gemini 3.8 Flash doubles to $1.50/$7.50 on January 1, 2027. If you are building a twelve-month cost model, use the later numbers, not the ones on the page today.
Caching and batching move the number more than model choice
A Claude cache hit costs 10% or less of the standard input price (2.5% on Fable 5.1), and the Batch API takes 50% off both directions; the two stack. A workload that reuses a large system prompt can land below a nominally cheaper model that you are calling uncached. Model selection is usually the third-largest lever, after caching and after not sending the context at all.
What the evidence says, and who is saying it
Almost every performance number in public circulation traces back to a vendor running its own harness. That does not make the numbers useless, but the attribution belongs next to the claim.
Arithmetic from Anthropic’s published list prices
- At list price, 10,000 support conversations of about 3,700 tokens each cost roughly $4 to $6 in tokens on Haiku 5.5.
- Haiku 5.5 cache reads are billed at 10% of the standard input rate, and the Batch API takes 50% off both directions.
This is an illustrative workload, split roughly 90% input and 10% output. Your token-per-conversation number is the variable that matters, and it is usually higher than you expect once policy documents and conversation history are in context.
What neither vendor publishes
- Deflection rate on your ticket mix.
- How often the model escalates when it should have answered, and answered when it should have escalated.
- CSAT on AI-handled tickets versus human-handled ones.
These are the only three numbers that decide whether support automation works, and no vendor can produce them for you. They come out of a shadow-mode run, which is why the section below exists.
Run it in shadow mode before a customer sees it
Support is the wrong place to learn from production. The good news is that support is unusually easy to test safely, because you have a backlog of resolved tickets with known correct answers and a human already graded them.
- Take 200 resolved tickets from the last quarter. Weight them the way your real queue is weighted, not toward the interesting ones. The boring repetitive tickets are where the ROI lives.
- Run both models against them with no customer in the loop. Same prompt, same retrieved context, same tools. Log the answer and the escalate/answer decision.
- Grade three things separately. Was the answer correct? Was it grounded in a real source passage? Did it make the right escalate-or-answer call? A model can score well on the first and badly on the third, and the third is what gets you in trouble.
- Count the confidently wrong answers specifically. Not the refusals — the answers that were fluent, plausible, and wrong. That number, not the average, is your risk. One wrong refund policy quote outweighs fifty good deflections.
- Price it on cost per resolved ticket. Include the escalated ones, which cost you tokens and a human. A model that is cheaper per token and escalates twice as often is not cheaper.
How to evaluate your agents has the test-set design, and hallucination prevention covers the grounding setup that keeps the confidently-wrong number down.
The honest bottom line
Pick Claude if the cost of a wrong answer is high — regulated industries, financial or medical information, anything where a customer could act on bad guidance. The citation behaviour and the readier escalation are worth real money in those settings, and the compliance surface is more developed.
Pick GPT-6 if your helpdesk already ships a first-class OpenAI integration and building the equivalent is a quarter of engineering time you do not have.
Route between them which is what most mature support deployments end up doing. Cheap model for tier-one deflection, stronger model for anything touching policy, money, or a frustrated customer. The routing rule is the product; the model is a component. See Claude for customer support for the full build.