Claude Opus 5.5
$80.00 / month at 10M input + 2M output tokens.
API pricing source ↗ · checked 2026-09-29
Base input/output rates; caching and batch tiers differ
HEAD TO HEAD
Pick two to four models. Compare category evidence, price, context, speed and capabilities side by side.
PRACTICAL COMPARISON
Check coding and reasoning separately, and review effort settings before choosing a model for a long workflow.
At 10 million input and 2 million output tokens per month, Claude Opus 5.5 has the lower listed token bill. Compare the same usage below.
$80.00 / month at 10M input + 2M output tokens.
API pricing source ↗ · checked 2026-09-29
Base input/output rates; caching and batch tiers differ
$200.00 / month at 10M input + 2M output tokens.
API pricing source ↗ · checked 2026-09-29
Base input/output rates; caching and batch tiers differ
Estimates cover standard input and output tokens. Caching, tools, retries, long-context rates and taxes can change the actual bill. Benchmark leads count each matched evaluation equally and describe the covered tasks.
INDEPENDENT BENCHMARK / SNAPSHOT 2026-10-16
This snapshot is over 14 days old. Review the dated benchmark records below for newer evidence.
Artificial Analysis Intelligence Index v4.3.2 · Higher is better. Selected models, sorted by this benchmark’s published score; equal scores are ties.
| Model | AA Intelligence Index | Tested effort | Evidence |
|---|---|---|---|
| Claude Opus 5.5 | 58 | Max, default fallback | Independent result ↗ |
This is a dated snapshot from one independent evaluator, on its own scale—not a Quantum Score or a complete market ranking. A missing Quantum Score means our broader evidence requirements are unmet; it does not mean the model is worse. Test dates are not published in these source summaries.
01 / SCORECARD
| Metric | Claude Opus 5.5 | Claude Fable 5.1 |
|---|---|---|
| Quantum Score | 88QUANTUM SCORE | 68QUANTUM SCORE |
| Confidence / coverage | Moderate confidence · 85% · 5 independent domains | Moderate confidence · 85% · 4 independent domains |
| Context window | 1,000,000 tokens | 1,000,000 tokens |
| Modalities / tools | text, image · tool use, vision | text, image · tool use, vision |
| Coding | 94 | 70 -24 vs first |
| Reasoning | 91 | 86 -5 vs first |
| Math | 92 | 99 +7 vs first |
REVAL™ⓘREVAL™ Research Evaluation Index — our proprietary 0–100 benchmark measuring real-world AI research performance. Methodology → | Not tested | Not tested |
| Agents | 96 | 70 -26 vs first |
| Multimodal | 98 | 52 -46 vs first |
| Speed | 90 tok/sMax · 2026-10-01 | Not measured |
| Value | 52 | 25 -27 vs first |
| Input / 1M | $4.00 | $10.00 |
| Output / 1M | $20.00 | $50.00 |
| Measured output speed | 90 tok/sMax · 2026-10-01 | Not measured |
THE DECISION
5 current independent evaluations use matching versions, methods, and test settings across every selected model. These show performance on the named tests; they do not establish a universal best model.
4.0 · Vals mini-SWE-agent avg@3; Anthropic fallbacks treated as failures
Vals 2026-09-27 · Vals 200 public tasks; raw pass rate; mini-SWE-agent
2026-06-25 · Public table; published model effort variant · LiveBench public table 2026-06-25
2026-06-25 · Public table; published model effort variant · LiveBench public table 2026-06-25
20261025 · LiveBench 2.0 published release table · Published model configuration; effort not specified in table
Different benchmarks test different tasks. Effort levels are shown because they can affect results and cost. Verification dates indicate when we checked a source; they are not necessarily evaluation dates.
Enter usage in millions of tokens. Estimates use the listed standard input and output prices in USD.
Lowest estimated cost among these models
Pricing source ↗ · verified 2026-09-29
Base input/output rates; caching and batch tiers differ
Pricing source ↗ · verified 2026-09-29
Base input/output rates; caching and batch tiers differ
This estimates API token charges. Subscription plans, caching, batch discounts, tools, taxes, and differences in tokens used per task can change the bill. Cost alone does not establish equivalent quality.
Listed capabilities: tool use, vision.
Independent benchmark records cover coding, math, reasoning, agents, multimodal. Coverage does not guarantee comparable tests for every selected model.
Listed capabilities: tool use, vision.
Independent benchmark records cover coding, math, agents, reasoning, multimodal. Coverage does not guarantee comparable tests for every selected model.
How we assess scores and missing evidence →
Create a free account to save these models and get alerts when their verified evidence or prices change.
Sign in to save ↗02 / EVIDENCE
| Benchmark | Claude Opus 5.5 | Claude Fable 5.1 |
|---|---|---|
| Terminal-Bench 4.0 Vals mini-SWE-agent avg@3 fallbacks failed · 4.0 · Vals mini-SWE-agent avg@3; Anthropic fallbacks treated as failures | 58.08 % independent lab ↗ | 50 % independent lab ↗ |
| ProofBench v1.1 Vals 100 proof tasks · 1.1 · Vals 100 proof tasks; ceiling effect at top scores | 100 % independent lab ↗ | 100 % independent lab ↗ |
| ProgramBench Vals 200 public tasks raw pass rate · Vals 2026-09-27 · Vals 200 public tasks; raw pass rate; mini-SWE-agent | 87 % independent lab ↗ | 82.7 % independent lab ↗ |
| LiveBench code_completion · 2026-06-25 · Public table; published model effort variant | 86.957 % benchmark maintainer ↗ | 82.609 % benchmark maintainer ↗ |
| LiveBench code_generation · 2026-06-25 · Public table; published model effort variant | 91.549 % benchmark maintainer ↗ | 90.141 % benchmark maintainer ↗ |
| LiveBench javascript · 2026-06-25 · Public table; published model effort variant | 81.818 % benchmark maintainer ↗ | 68.182 % benchmark maintainer ↗ |
| LiveBench AMPS_Hard · 2026-06-25 · Public table; published model effort variant | 98 % benchmark maintainer ↗ | 99 % benchmark maintainer ↗ |
| LiveBench olympiad · 2026-06-25 · Public table; published model effort variant | 93.251 % benchmark maintainer ↗ | 92.968 % benchmark maintainer ↗ |
| LiveBench theory_of_mind · 2026-06-25 · Public table; published model effort variant | 84.615 % benchmark maintainer ↗ | 80.769 % benchmark maintainer ↗ |
| LiveBench zebra_puzzle · 2026-06-25 · Public table; published model effort variant | 100 % benchmark maintainer ↗ | 100 % benchmark maintainer ↗ |
| MMMU-Pro Mercor public leaderboard · MMMU-Pro Mercor public set · Mercor independent evaluation | 87.4 % independent lab ↗ | 83.9 % independent lab ↗ |
| MedXpertQA MM Mercor public leaderboard · MedXpertQA MM Mercor 200 public tasks · Mercor independent evaluation | 85.7 % independent lab ↗ | 79.7 % independent lab ↗ |
| MedXpertQA MM Mercor 51 held-out tasks · MedXpertQA MM Mercor Extended 51 held-out tasks · Mercor independent evaluation | 71.2 % independent lab ↗ | 66 % independent lab ↗ |
| LiveBench 2.0 coding · 20261025 · LiveBench 2.0 published release table | 96.3 % benchmark maintainer ↗ | 97.7 % benchmark maintainer ↗ |
| LiveBench 2.0 reasoning · 20261025 · LiveBench 2.0 published release table | 92.2 % benchmark maintainer ↗ | 91.7 % benchmark maintainer ↗ |
| LiveBench 2.0 long_agentic · 20261025 · LiveBench 2.0 published release table | 72.8 % benchmark maintainer ↗ | 77.9 % benchmark maintainer ↗ |
AS THE EVIDENCE CHANGED
Dated releases, price events, and benchmark publications. A price or test change can alter the factual comparison; this timeline does not pick a winner.
2026-11-07 · Claude Fable 5.1 recorded coding score changed ↗ · source
2026-10-02 · Claude Fable 5.1 recorded coding score changed ↗ · source
2026-10-02 · Claude Opus 5.5 recorded coding score changed ↗ · source
2026-10-01 · Claude Fable 5.1 recorded coding score changed ↗ · source
2026-10-01 · Claude Opus 5.5 recorded coding score changed ↗ · source
2026-09-22 · Claude Opus 5.5 released ↗ · source
2026-09-01 · Claude Fable 5.1 released ↗ · source
2026-11-07 verified (test date not published) · claude-fable-5-1 entered LiveBench 2.0 coding (20261025) at 97.7 % · source ↗
2026-11-07 verified (test date not published) · claude-fable-5-1 entered LiveBench 2.0 reasoning (20261025) at 91.7 % · source ↗
2026-11-07 verified (test date not published) · claude-fable-5-1 entered LiveBench 2.0 long_agentic (20261025) at 77.9 % · source ↗
2026-11-07 verified (test date not published) · claude-opus-5-5 entered LiveBench 2.0 coding (20261025) at 96.3 % · source ↗
2026-11-07 verified (test date not published) · claude-opus-5-5 entered LiveBench 2.0 reasoning (20261025) at 92.2 % · source ↗
2026-11-07 verified (test date not published) · claude-opus-5-5 entered LiveBench 2.0 long_agentic (20261025) at 72.8 % · source ↗
2026-10-17 verified (test date not published) · claude-fable-5-1 entered LiveBench code_completion (2026-06-25) at 82.609 % · source ↗
2026-10-17 verified (test date not published) · claude-fable-5-1 entered LiveBench code_generation (2026-06-25) at 90.141 % · source ↗
2026-10-17 verified (test date not published) · claude-fable-5-1 entered LiveBench javascript (2026-06-25) at 68.182 % · source ↗
2026-10-17 verified (test date not published) · claude-fable-5-1 entered LiveBench AMPS_Hard (2026-06-25) at 99 % · source ↗
2026-10-17 verified (test date not published) · claude-fable-5-1 entered LiveBench olympiad (2026-06-25) at 92.968 % · source ↗
2026-10-17 verified (test date not published) · claude-fable-5-1 entered LiveBench theory_of_mind (2026-06-25) at 80.769 % · source ↗
An overall score appears only after enough current, comparable, independently sourced results exist. Pricing and individual benchmark results remain visible meanwhile.