#42 Claude Opus 4.6
medium
Invalid SVG
- Cost
- $0.000
- Time
- 300.0s
- Tokens
- 0 tok
Anthropic
Release: 2026-02-05
Tested on: 2026-04-11 01:44
anthropic/claude-opus-4.6::medium
~5T total (~500B active)MoEClosedEstimated Summary
Claude Opus 4.6 scores 7.6 on AI BENCHY and ranks #42. It has N/A reliability, a 70.4% pass rate, $1.446 total cost, and 21.08s average response time.
What makes Claude Opus 4.6 unique: It stands out most in General Intelligence, where it ranks #1, while Anti-AI Tricks is its weakest area at #18.
Model facts
Researched on 2026-08-12
- Parameters
- ~5T total (~500B active)
- Architecture
- MoE
- Availability
- Closed
- License
- -
Best estimate from public evidence; the vendor did not disclose every value. External industry estimate; vendor has not disclosed the architecture.
7.6
Consistency
9.1
N/A
$1.446
Total Output Tokens
48,767
Total Input Tokens
0
Input Price
$5.000 / 1M
Output Price
$25.000 / 1M
Flaky tests
2
Flaky tests had mixed outcomes across runs (at least one pass and one fail).
Hamster playing table tennis
Prompt: Create a detailed SVG illustration of a hamster playing table tennis.
Run history
| Tested on | Score | Reliability | Tests Correct | Total Cost | Compare |
|---|---|---|---|---|---|
| 2026-08-14 02:08 Re-test | 7.7 | 10.0 | $2.961 | Compare | |
| 2026-07-16 22:31 New test added | 7.7 | 10.0 | $3.059 | Compare | |
| 2026-06-04 13:45 New test added | 7.0 | 10.0 | $2.053 | Compare | |
| 2026-05-22 00:17 Suite changed | 7.2 | 10.0 | $1.896 | Compare | |
| 2026-04-11 01:44 First recorded run | 7.6 | N/A | $1.446 | Current run |
Run comparison
| Run | Benchmark coverage | Score | Consistency | Reliability | Tests Correct | Flaky tests | Total Output Tokens | Total Input Tokens | Total Cost | Response Time (avg) |
|---|---|---|---|---|---|---|---|---|---|---|
| 2026-04-11 01:44 · First recorded run | 54/54 attempts | 7.6 | 9.1 | N/A | 12/18 | 2 | 48,767 | 0 | $1.446 | 21.08s |
| 2026-05-22 00:17 · Suite changed | 60/60 attempts | 7.2 | 9.1 | 10.0 | 12/20 | 2 | 65,814 | 0 | $1.896 | 25.44s |
| Difference | — | +0.4 | 0.0 | 0 | 0 | -17047 | 0 | -$0.451 | -4362ms |
Benchmark coverage differs: 54/54 attempts (Target: 3 repeats per test) versus 60/60 attempts (Target: 3 repeats per test). Totals and repeat-sensitive metrics are not directly comparable.
These two runs used different benchmark suites, so the deltas reflect both model changes and suite changes.
Charts
Choose the first model, then click a second model to open a side-by-side page.
Top Models by Score
Score vs Total Cost
Response Time (avg)
Score vs Response Time (avg)
Total Output Tokens
Score vs Total Output Tokens
Quick Compare
Claude Opus 4.6mediumvsSeed-2.0-CodelowClaude Opus 4.6mediumvsDeepSeek V4 ProhighClaude Opus 4.6mediumvsGLM 5mediumClaude Opus 4.6mediumvsDeepSeek V4 Flash 0423highClaude Opus 4.6mediumvsQwen3.7 FlashhighClaude Opus 4.6mediumvsGPT-5.6 LunahighClaude Opus 4.6mediumvsClaude Opus 4.8lowClaude Opus 4.6mediumvsSolar Pro 4xhighClaude Opus 4.6mediumvsSeed-2.0-LitemediumClaude Opus 4.6mediumvsNemotron 3 UltramediumFree Available
Category Breakdown
| Category | Score | Consistency | Tests Correct |
|---|---|---|---|
| Anti-AI Tricks | 6.4 | 5.8 | |
| Coding | 10.0 | 10.0 | |
| Combined | 10.0 | 10.0 | |
| Data parsing and extraction | 10.0 | 10.0 | |
| Domain specific | 3.0 | 10.0 | |
| General Intelligence | 10.0 | 10.0 | |
| Instructions following | 10.0 | 10.0 | |
| Puzzle Solving | 7.7 | 10.0 | |
| Tool Calling | 10.0 | 10.0 |