Claude Opus 4.8 pricing: real cost per million tokens
What Anthropic's Claude Opus 4.8 really costs across 3 plans from 2 providers: monthly fees, usable allowances and real per-token prices, against the official API list price and independent leaderboards.
Data updated · How real prices are computed
As of 2026-10-08, the cheapest way to use Claude Opus 4.8 is Claude Pro at a real cost of about $0.00772 per million tokens ($20/month for roughly 2.589B tokens a month). We track 3 plans and API channels that serve Claude Opus 4.8. Official API list price per million tokens: $0.5 cached input, $5 input, $25 output — about $0.735 per million tokens on a standard workload mix.
- Lowest real price
- $0.00772/MTokClaude Pro
- API list price
- $0.735/MTokstandard workload mix
- Most monthly tokens
- 15.7BClaude Max 20x (9/14+)
- Plans & APIs
- 33 subscriptions · 0 metered APIs
Plot Claude Opus 4.8 on the price–capability chart →
Claude Opus 4.8 plans and real prices
Real price = monthly fee ÷ tokens you can actually use in a month. “vs API” shows how many times cheaper than the official API list price (standard workload mix) the plan is. “Evidence” opens the full calculation and sources for that row.
| Plan | Provider | Billing | Monthly fee | Monthly tokens | Real $/MTok | vs API | Confidence | Data date | |
|---|---|---|---|---|---|---|---|---|---|
| Claude Pro | Anthropic | Subscription | $20/mo | 2.589B | $0.00772 | 95.1× | Low | Evidence ↓ | |
| Claude Max 20x (9/14+) | Anthropic | Subscription | $200/mo | 15.7B | $0.01274 | 57.7× | Low | Evidence ↓ | |
| Droid Max | Factory | Subscription | $200/mo | 2.021B | $0.09897 | 7.4× | Medium | Evidence ↓ |
Behind each number
For every plan: the data date, the exact calculation, the quota basis, how leaderboard scores were mapped, and the original adoption evidence with its sources.
Claude ProAnthropic$0.00772499/MTok
- Plan
- Claude Pro · Anthropic
- Data date
- 2026-09-14 – 2026-09-22 · derived; uses the data date of Claude Pro · claude-opus-5
- Real price / MTok
- $0.00772499
- Monthly tokens
- 2,589,000,000 (2.589B)
- Quota confidence
- Low
- Calculation
- $20 / month ÷ 2,589,000,000 tokens × 1,000,000 ≈ $0.00772499 / MTok
- Quota basis
Anthropic workload: 97% cache reads / 2.5% cache writes / 0.5% output
Official API list (cached / in / out, per MTok): $0.5 / $5 / $25 · standard-load blended ≈ $0.735/MTok
- Benchmark mapping
- Artificial Analysis Intelligence Index v4.3: Medium (1 configuration)
Terminal-Bench 4.0: Low (1 configuration)
Terminal-Bench 4.0 (AA): Low (1 configuration)
Code Arena | WebDev Overall: Low (2 configurations)
Agent Arena · Overall (Ranking / Pareto): Low (1 configuration)
DeepSWE v1.1: Low (5 configurations)Mapping confidence: how directly each score refers to the model served here.
Adoption evidence · source and rationale (translated from the Chinese record)
Derived from same-plan claude-opus-5 2.589B × 1.0
Same price and pool as Opus 5; 2.762B→2.589B derived from the Opus 5 normalization baseline (the earlier 18.78 was the round8 panel reading, 27.62 the #53 raw); the historical 1.588B measurement is kept as pre-round5 evidence, not overwritten; community-issues-round1-2026-09-29.json, claude-adoption-round8-2026-09-20.json; workload-conversion-round1-2026-09-30.json
claude-adoption-round8-2026-09-20.jsoncommunity-issues-round1-2026-09-29.jsonworkload-conversion-round1-2026-09-30.jsonBrowse the public evidence archive
Claude Max 20x (9/14+)Anthropic$0.0127389/MTok
- Data date
- 2026-08 · derived; uses the data date of Claude Max 20x (9/14+) · claude-opus-5
- Real price / MTok
- $0.0127389
- Monthly tokens
- 15,700,000,000 (15.7B)
- Quota confidence
- Low
- Calculation
- $200 / month ÷ 15,700,000,000 tokens × 1,000,000 ≈ $0.0127389 / MTok
- Quota basis
Not workload-normalized: sample lacks a token breakdown (or is an official absolute token table); raw tokens used as-is
Official API list (cached / in / out, per MTok): $0.5 / $5 / $25 · standard-load blended ≈ $0.735/MTok
- Benchmark mapping
- Artificial Analysis Intelligence Index v4.3: Medium (1 configuration)
Terminal-Bench 4.0: Low (1 configuration)
Terminal-Bench 4.0 (AA): Low (1 configuration)
Code Arena | WebDev Overall: Low (2 configurations)
Agent Arena · Overall (Ranking / Pareto): Low (1 configuration)
DeepSWE v1.1: Low (5 configurations)Mapping confidence: how directly each score refers to the model served here.
Adoption evidence · source and rationale (translated from the Chinese record)
Derived from same-plan claude-opus-5 15.7B × 1.0
Old 8B→15.7B; same price as Opus 5, derived on the 9/14 permanent baseline; claude-adoption-round6-2026-09-06.json
claude-adoption-round6-2026-09-06.jsonBrowse the public evidence archive
Droid MaxFactory$0.0989707/MTok
- Data date
- 2026-09-29 · derived; uses the data date of Droid Max · claude-opus-5.5
- Real price / MTok
- $0.0989707
- Monthly tokens
- 2,020,800,000 (2.021B)
- Quota confidence
- Medium
- Calculation
- $200 / month ÷ 2,020,800,000 tokens × 1,000,000 ≈ $0.0989707 / MTok
- Quota basis
Anthropic workload: 97% cache reads / 2.5% cache writes / 0.5% output
Official API list (cached / in / out, per MTok): $0.5 / $5 / $25 · standard-load blended ≈ $0.735/MTok
- Benchmark mapping
- Artificial Analysis Intelligence Index v4.3: Medium (1 configuration)
Terminal-Bench 4.0: Low (1 configuration)
Terminal-Bench 4.0 (AA): Low (1 configuration)
Code Arena | WebDev Overall: Low (2 configurations)
Agent Arena · Overall (Ranking / Pareto): Low (1 configuration)
DeepSWE v1.1: Low (5 configurations)Mapping confidence: how directly each score refers to the model served here.
Adoption evidence · source and rationale (translated from the Chinese record)
Derived from same-plan claude-opus-5.5 2.526B × 0.8
Factory official multipliers Opus 5.5 1.6× / claude-opus-4.8 2×: the same Standard Usage pool scaled by multiplier — not a measurement of this model; inherits the Opus 5.5 row's Anthropic-tier workload, not recomputed for each model's own price gap; docs.factory.ai/docs/models; droid-model-multipliers-2026-09-29.json
droid-model-multipliers-2026-09-29.jsonBrowse the public evidence archive
Claude Opus 4.8 benchmark scores
Highest archived configuration per leaderboard; ranks are among all models each board archives in this dataset. Scores are never mixed across boards.
| Leaderboard | Metric | Score | Rank | Configuration | Effort | Harness | Configs | Snapshot |
|---|---|---|---|---|---|---|---|---|
| Artificial Analysis Intelligence Index v4.3 | Intelligence Index | 41.79 | #20 / 60 | Claude Opus 4.8 (max) | Max | — | 1 | 2026-10-01 |
| Terminal-Bench 4.0 | Resolution Rate % | 23.64 | #11 / 18 | Claude Code - Opus 4.8 (max) | Max | Claude Code | 1 | 2026-09-21 |
| Terminal-Bench 4.0 (AA) | Resolution Rate % | 21.72 | #20 / 53 | Artificial Analysis - Claude Opus 4.8 (max) | Max | Artificial Analysis | 1 | 2026-10-01 |
| Code Arena | WebDev Overall | Arena Score | 1,556 | #29 / 54 | claude-opus-4-8-high | High | — | 2 | 2026-09-30 |
| Agent Arena · Overall (Ranking / Pareto) | Net Improvement % | 6.92 | #8 / 39 | Claude Opus 4.8 (High) | High | — | 1 | 2026-09-30 |
| DeepSWE v1.1 | Pass@1 % | 58.97 | #19 / 33 | mini-swe-agent - claude-opus-4-8 (max) | Max | mini-swe-agent | 5 | 2026-09-03 |
Every archived configuration
Artificial Analysis Intelligence Index v4.3 1 configuration
Terminal-Bench 4.0 1 configuration
| Configuration | Score | Score interval | Harness / effort / mode | Source task cost (mean / median) | Checked | Links |
|---|---|---|---|---|---|---|
| Claude Code - Opus 4.8 (max)Official Terminal-Bench 4.0 leaderboard row; score belongs to this harness × model × effort configuration. | 23.64 | 20.08 – 27.2 | Claude Code / Max / — | — / — | 2026-10-01 | Source · Archive |
Terminal-Bench 4.0 (AA) 1 configuration
| Configuration | Score | Score interval | Harness / effort / mode | Source task cost (mean / median) | Checked | Links |
|---|---|---|---|---|---|---|
| Artificial Analysis - Claude Opus 4.8 (max)AA's own-harness independent measurement of Terminal-Bench 4.0 (the terminalBench40 field in the models-board payload, fraction×100); scored separately from the official tbench.ai board. | 21.7172 | — | Artificial Analysis / Max / — | — / — | 2026-10-01 | Source · Archive |
Code Arena | WebDev Overall 2 configurations
| Configuration | Score | Score interval | Harness / effort / mode | Source task cost (mean / median) | Checked | Links |
|---|---|---|---|---|---|---|
| claude-opus-4-8-high | 1,556 | 1,550.3 – 1,561.7 | — / High / — | — / — | 2026-10-01 | Source · Archive |
| claude-opus-4-8 | 1,534 | 1,528.3 – 1,539.7 | — / — / — | — / — | 2026-10-01 | Source · Archive |
Agent Arena · Overall (Ranking / Pareto) 1 configuration
DeepSWE v1.1 5 configurations
| Configuration | Score | Score interval | Harness / effort / mode | Source task cost (mean / median) | Checked | Links |
|---|---|---|---|---|---|---|
| mini-swe-agent - claude-opus-4-8 (max) | 58.9744 | 57.21 – 60.739 | mini-swe-agent / Max / — | $13.22 / $12.35 | 2026-09-12 | Source · Archive |
| mini-swe-agent - claude-opus-4-8 (xhigh) | 54.3624 | 50.649 – 58.076 | mini-swe-agent / xhigh / — | $8.006 / $7.049 | 2026-09-12 | Source · Archive |
| mini-swe-agent - claude-opus-4-8 (high) | 51.7699 | 47.208 – 56.332 | mini-swe-agent / High / — | $4.282 / $3.664 | 2026-09-12 | Source · Archive |
| mini-swe-agent - claude-opus-4-8 (medium) | 48.6726 | 46.433 – 50.912 | mini-swe-agent / Medium / — | $3.444 / $2.84 | 2026-09-12 | Source · Archive |
| mini-swe-agent - claude-opus-4-8 (low) | 40.7982 | 39.335 – 42.262 | mini-swe-agent / Low / — | $2.293 / $1.824 | 2026-09-12 | Source · Archive |
These are reference mappings, not benchmarks of a subscription or API channel. Harness and quota-measurement effort alignment are unverified. Source task costs are not subscription task costs.
- Exact served-model reference; quota-measurement effort is unverified.
- Exact served-model reference only; product harness and quota-measurement effort are unverified. Not a benchmark measurement of this subscription or API channel.
Claude Opus 4.8 official API pricing
| Cached input | Input | Output |
|---|---|---|
| $0.5 | $5 | $25 |
Frequently asked questions
How much does Claude Opus 4.8 really cost per million tokens?
Measured as monthly fee ÷ tokens you can actually use, Claude Opus 4.8 ranges from $0.00772 (Claude Pro) to $0.09897 (Droid Max) per million tokens.
Which subscription gives the most Claude Opus 4.8 usage?
Claude Max 20x (9/14+) allows roughly 15.7B tokens of Claude Opus 4.8 per month for $200.
What is the official API price of Claude Opus 4.8?
Official API list price per million tokens: $0.5 cached input, $5 input, $25 output — about $0.735 per million tokens on a standard workload mix.
How does Claude Opus 4.8 score on independent leaderboards?
Claude Opus 4.8 has scores on 6 leaderboards; its strongest placement is 6.92 (Net Improvement %) on Agent Arena · Overall (Ranking / Pareto), #8 of 39 models tracked.
Who makes Claude Opus 4.8, and where can I use it?
Claude Opus 4.8 (model ID claude-opus-4.8) is made by Anthropic and is available through Anthropic, Factory.
More Anthropic models
| Model | Lowest real $/MTok | Cheapest plan | Plans |
|---|---|---|---|
| Claude Haiku 5.5 | $0.00028 | Claude Max 5x (9/14+) | 4 |
| Claude Sonnet 5.5 | $0.00273 | Claude Max 5x (9/14+) | 6 |
| Claude Sonnet 5 | $0.00309 | Claude Pro | 4 |
| Claude Opus 5.5 | $0.00545 | Claude Max 5x (9/14+) | 9 |
| Claude Opus 5 | $0.00772 | Claude Pro | 5 |
| Claude Fable 5 | $0.1083 | Claude Max 5x (9/14+) | 4 |
| Claude Fable 5.1 | $0.1157 | Claude Max 20x (9/14+) | 4 |