AI Codex
CompareClaude vs GPT-5.6 for Customer Support

Comparison · verified August 21, 2026

Claude vs GPT-5.6 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

Claude
GPT-5.6

The tier you would actually run

Haiku 4.5 at $1 / $5 for tier-one deflection; Sonnet 5 at $2 / $10 when the answer needs reasoning over policy. Anthropic’s own worked example puts 10,000 support conversations at roughly $37 on Haiku 4.5.

Luna at $0.20 / $1.20 is the cheapest credible option on the market for high-volume deflection. Terra at $2 / $12 for harder tickets.

Cost at 100k tickets/month

On Haiku 4.5, in the region of $370/month in tokens at ~3,700 tokens per conversation, before caching. Cache hits cost 10% of input, and your policy documents are the same on every ticket — so real-world cost lands well below the arithmetic.

Luna is roughly a fifth of Haiku 4.5’s input rate, so the token line is materially cheaper. Check whether your context length crosses OpenAI’s long-context threshold, which doubles the rate.

Grounding answers in your help centre

The Citations API returns the exact source passage behind a claim. For support that is not a nicety — it is how an agent tells the difference between quoting your refund policy and inventing one.

Retrieval and file search are available; verify what the citation guarantees are for your setup rather than assuming parity.

Escalation and refusal behaviour

Constitutional AI training produces a model that declines and escalates comparatively readily. In support that is usually the behaviour you want; in sales it can read as unhelpful.

Generally more willing to attempt an answer. Better completion rates, and a larger surface for confidently wrong answers reaching a customer.

Compliance and audit trail

Compliance API returns activity, chats, files, projects, and since August 11, 2026 the transcripts of Cowork and Claude Code sessions running on employees’ own machines. Inference hooks can deny a prompt before the model sees it.

Enterprise admin and logging available through OpenAI and Azure. Map the specific controls against your retention and DLP requirements — the surfaces are not equivalent.

Where the model runs

Claude API, Bedrock, Claude Platform on AWS, Google Cloud, Microsoft Foundry. `inference_geo` pins the jurisdiction at a 1.1x multiplier.

OpenAI API and Azure, with Azure data-zone options.

Helpdesk integration

Connectors and MCP for Intercom, Zendesk, and similar. Claude Tag puts Claude in Slack as a tagged participant for internal escalation threads.

Broader catalogue of off-the-shelf helpdesk integrations, and more vendors ship an OpenAI connector by default.

What each one costs

Last verified August 21, 2026

USD per million tokens, standard rates, no discounts applied. Read the three caveats underneath before you put these numbers in a spreadsheet.

ModelTierInputOutputNotes
Claude Fable 5frontier$10$50The frontier model. Twice the price of Opus 5 for the hardest work.
Claude Opus 5frontier$5$25The default flagship. Thinking on by default. Fast mode available at $10/$50.
Claude Sonnet 5balanced$2$10What most production agent fleets run on. The September 2026 increase to $3/$15 was cancelled.
Claude Haiku 4.5fast$1$5High-volume and latency-sensitive work.
GPT-5.6 Solfrontier$5$30The flagship tier. Long-context requests meter at $10/$45.
GPT-5.6 Terrabalanced$2$12The everyday tier, cut to this rate on July 30, 2026. Long-context meters at $4/$18.
GPT-5.6 Lunafast$0.20$1.2The cheapest frontier-family option on the market. Long-context meters at $0.40/$1.80.

Anthropic changed tokenizers at Claude 4.7

Claude 4.7 and later — including Opus 5, Sonnet 5, and Fable 5 — 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.

OpenAI meters long context separately

GPT-5.6 publishes a second, higher rate for long-context requests: Sol goes from $5/$30 to $10/$45, Terra from $2/$12 to $4/$18, Luna from $0.20/$1.20 to $0.40/$1.80. Anthropic includes the full 1M-token window at standard pricing on Claude 4.6 and later — a 900k-token request costs the same per token as a 9k one. If your workload is context-heavy, that difference is larger than the headline gap.

Two of Google’s current rates are introductory

Gemini 3.7 Flash and 3.6 Flash are priced at $0.75/$3.75 only through December 31, 2026. Both double to $1.50/$7.50 in 2027. If you are building a twelve-month cost model on Gemini Flash, model the 2027 number, not the one on the page today.

Caching and batching move the number more than model choice

A Claude cache hit costs 10% of the standard input price, 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.

Anthropic, published cost guidance

  • A worked example of 10,000 support conversations on Haiku 4.5 at roughly 3,700 tokens each totals about $37 in tokens.
  • Cache reads are billed at 10% of the standard input rate, and the Batch API takes 50% off both directions.

These are Anthropic’s own figures for an illustrative workload. 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.

  1. 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.
  2. 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.
  3. 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.
  4. 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.
  5. 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-5.6 if you are running very high volume on low-stakes tickets, where Luna at $0.20 / $1.20 changes the unit economics outright, or 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.

Related

Claude for customer support

The full build — routing, grounding, escalation.

CS team playbook

Running a support team with AI in the loop.

Document analysis

The comparison for contracts and long documents.

Cost calculator

Model spend at your ticket volume.