AI Model Reviews

Claude Opus 5.5 Review: Anthropic's New AI Workhorse

Claude Opus 5.5 is not interesting simply because it tops a launch table. It is interesting because Anthropic is trying to turn frontier-level performance into an everyday workhorse: cheaper per token, more efficient per task, faster to respond, and easier to supervise across long agent sessions.

The short answer: Opus 5.5 is the strongest default Claude model for difficult coding and knowledge work. Start with medium effort, measure cost per accepted task, and escalate only when higher effort improves the result. Keep Fable 5.1 for planning, security review, and high-consequence work. Keep GPT-6 Astra in the comparison set for scientific tasks, computer control, and workflows where your own tests favor it.

Watch Matthew Berman's Review and Interview

Source and credit: Anthropic went CRAZY (Opus 5.5), published by Matthew Berman on 22 September 2026. The second half is an interview with Thariq, a member of Anthropic's technical staff. Forward Future collected its wider Opus 5.5 tests and review resources here.

Why Opus 5.5 Looks Like the Workhorse Model

Anthropic's positioning is unusual. It says Opus 5.5 performs at Fable 5.1's level on most work, yet is priced below Opus 5 and needs less compute to serve. In the interview, Thariq calls it a strong daily driver and says he used it across an entire personal-site rebuild.

He gives Fable 5.1 a narrower role: discrete planning, code review, security analysis, and tasks where the cost of a wrong answer is high enough that token price matters less. That division is more useful than a single leaderboard rank:

ModelOperational roleUse it when
Claude Opus 5.5Daily workhorseCoding agents, migrations, audits, research, and knowledge work
Claude Fable 5.1High-consequence specialistPlanning, security review, code review, and expensive-to-reverse decisions
GPT-6 AstraCross-vendor challengerScientific work, demanding computer use, and tasks where your tests show an advantage

What the Benchmark Table Actually Shows

Anthropic reports 66.4% on Terminal-Bench 4.0, 54.4% on FrontierCode 1.1, 57.8% on CursorBench 4.0, and 1846 Elo on GDPval-AA 2.1. Those results place Opus 5.5 ahead of Fable 5.1 and Opus 5 in the published table.

GPT-6 Astra remains ahead on two listed tasks: AutomationBench at 41.4% versus 40.0%, and Terminal-Bench-Science at 64.6% versus 58.7%. Several computer-use and chart-reading rows do not include Astra, so they cannot support a direct win over it.

The footnotes matter. Opus results use different effort settings across tests. Safeguard interventions can trigger fallback models, while Zapier's AutomationBench run counted interventions as failures. Anthropic also says benchmark margins at this capability level are becoming less reliable and that the real-world gap between Opus 5.5 and Fable 5.1 is narrower than the headline scores suggest.

Evidence label: these are vendor-reported launch evaluations. They are useful for choosing what to test, not sufficient for changing a production route by themselves.

Cost per Task Is the Real Pricing Story

Standard API prices are $4 per million input tokens, $20 per million output tokens, $0.20 per million cache reads, and $5 per million cache writes. That is a 20% input-and-output reduction from Opus 5, while cache reads fall by 60%. Anthropic says fewer tokens and tool calls increase the typical workload saving to about 40%.

Matthew's central point is correct: token price is only one component. An agent that takes ten attempts, repeats context, or writes a long explanation can cost more than a nominally expensive model that finishes cleanly.

Track this instead:

Accepted-result cost = model tokens + cache + tool fees + retries + elapsed compute + reviewer time, divided by outputs that meet the definition of done.

Anthropic's launch examples support the efficiency thesis, although they remain vendor-selected evidence. It reports an HAProxy migration finishing in 9.5 hours versus 12 for Fable 5.1 at 51% lower cost, and a merger-analysis workbook and presentation finishing in 63 minutes versus 93 for Opus 5 at half the cost.

Max Effort Is Not Automatically the Best Choice

The quality-versus-cost charts contain a practical lesson that gets lost in the launch score. On FrontierCode, the medium setting can score better than a much more expensive max run. On Terminal-Bench, medium or high sits closer to the useful upper-left corner: high acceptance at a lower task cost.

More thinking can create diminishing returns, longer latency, and occasionally worse benchmark performance. The safest default is:

  1. Start at medium. Use a representative task with a strict acceptance check.
  2. Escalate failures to high. Record whether the extra spend fixes the actual failure mode.
  3. Reserve max for proven cases. Use it only when repeated evaluation shows a material lift.
  4. Route simple work downward. Classification, extraction, and short edits usually belong on a cheaper model.

Five Useful Things Anthropic Revealed

1. Claude helps build Claude incrementally

Thariq describes recursive self-improvement as many small gains rather than one autonomous model redesigning itself. Claude writes code, helps with research and evaluation, and contributes to the next system as part of an engineering process.

2. Pacing the frontier does not mean freezing releases

His explanation separates unreleased frontier capabilities from models Anthropic believes it can deploy with robust classifiers, fallbacks, and external evaluation. Opus 5.5 is presented as bringing previously expensive capability to a wider audience, not as the unreleased edge that the pacing proposal targets.

3. Better models can expose an outdated harness

The Claude Code team recently removed parts of its system prompt because old instructions can constrain a more capable model. Anthropic also introduced plugin evaluation so teams can check whether skills and plugins still help the newest model instead of forcing obsolete behavior.

4. Concise communication is operational

Shorter summaries are not merely style polish. When a person supervises 10 or 20 agents, every long status report adds context-switching cost. Opus 5.5 is designed to put the result first and follow explicit writing rules more closely.

5. Tool calling is part of the capability gain

Thariq highlights stronger tool use and custom agent harnesses. That matters because production value comes from reliable actions, verification, and handoffs, not just a better answer inside chat.

What Still Needs Work

The interview is unusually useful because it names unfinished areas. Thariq says combined browser and computer use still needs work so the product can express the capability shown in benchmarks. He also points to personalization, memory, and knowing when to challenge the user without becoming intrusive.

Safety is another boundary. Anthropic says Opus 5.5 achieved its strongest automated behavioral-audit result and improved prompt-injection resistance. Because it approaches Mythos 5.1 in biology and cybersecurity, it ships with Fable-like safeguards and verification programs for vetted research and security organizations. Those controls can affect availability and benchmark behavior.

Finally, the video is launch analysis plus an Anthropic interview, not an independent controlled benchmark. Matthew is enthusiastic, and the guest works for the vendor. Use the official results and interview as evidence about positioning and design decisions; use your own repeated trials to make a purchasing decision.

A Practical Routing Policy

TaskStart withEscalate when
Large code migration or auditOpus 5.5 mediumAcceptance tests fail or architectural judgment is weak
Daily coding-agent workOpus 5.5 mediumHigh materially reduces retries on your suite
Planning and security reviewFable 5.1A second model is required for independent review
Browser and computer automationOpus 5.5 and Astra side by sideChoose from completion rate, interventions, and recovery quality
Routine extraction or classificationA cheaper modelConfidence is low or the decision carries material risk

A Five-Run Evaluation Before You Switch

  1. Pick five recent tasks your current model completed, including at least one failure.
  2. Freeze the files, tools, permissions, prompt, and definition of done.
  3. Run Opus 5.5 at medium and high, then compare with your current route.
  4. Measure completion, elapsed time, tokens, tool calls, retries, and reviewer corrections.
  5. Adopt it only where its accepted-result cost is lower or the quality gain justifies the difference.

For more examples beyond this interview, use Forward Future's Opus 5.5 test collection and the site's companion four-build blind comparison.

Video Chapters

TimeChapterTimeChapter
00:00Introducing Opus 5.502:10Performance benchmarks
06:13Pricing and efficiency06:44Speed improvements
06:52Cost per task07:16Quality versus cost
11:55Communication and UI12:03Safety and alignment
13:03Model scalability16:26Independent intelligence index
17:04Anthropic interview29:50Harnesses, skills, and tool use

Sources and Links

Common questions

Is Claude Opus 5.5 better than Fable 5.1?
Anthropic reports higher Opus 5.5 scores on several coding and knowledge-work benchmarks, but says the real-world gap is narrower than the table suggests. In the interview, Anthropic positions Opus 5.5 as the daily driver and Fable 5.1 as a specialist for planning, security review, and high-consequence tasks.
How much does Claude Opus 5.5 cost?
Standard API pricing is $4 per million input tokens, $20 per million output tokens, $0.20 per million cache-read tokens, and $5 per million cache-write tokens. Fast mode costs $8 input and $40 output per million tokens.
Why is cost per task more useful than token price?
Two models with similar token prices can use very different numbers of tokens, steps, retries, and tool calls. Cost per accepted task captures the complete run and is a better operational measure than the public rate alone.
Which thinking effort should I use with Opus 5.5?
Start at medium for routine coding and knowledge work, then escalate only when a frozen evaluation shows higher effort improves acceptance. In Anthropic's charts, medium or high sometimes produced a better quality-cost tradeoff than max.
What did the Anthropic interview add?
Thariq described Opus 5.5 as an everyday model, kept Fable 5.1 for discrete planning and security tasks, explained that Claude incrementally helps build Claude, and said browser use, personalization, memory, and harness design still need improvement.
Where can I see Forward Future's Opus 5.5 tests?
Forward Future published its test collection and model-review resources at forwardfuture.com/opus-5-5-review. The linked page is the best companion to the video for current examples and follow-up testing.
Share
X LinkedIn Reddit
Build Yours

Want a system
like this one?

Book a free 30-minute call. We map your situation, identify the highest-impact automation, and figure out if we are a fit.

Book Free 30-min Call