Claude-Haiku-3
Legacy model retained for compatibility with older integrations
Shogo Agent Index
UnrankedNot ranked yet: 0 agentic benchmarks measured, and we need at least 2 plus an overall capability score. It will be ranked as results are published.
Agentic coding
7
estimated
Long-horizon autonomy
15
estimated
Knowledge-work agents
9
estimated
Computer use
1
estimated
Security & context learning
5
estimated
Price
- Input
- $0.21/M
- Output
- $1.10/M
- Cache read
- $0.021/M
- 3:1 blend
- $0.43/M
- Agent-effective
- $0.10/M
Agent-effective = 20 input tokens per output token, 85% cache hits. Why.
Agentic benchmarks
| Benchmark | Score | Normalised | Index weight | Harness / source |
|---|---|---|---|---|
| SWE-bench Verified | est. | 8% | predicted 25 | |
| Terminal-Bench 2.0 | est. | 8% | predicted 1 | |
| DeepSWE | est. | 6% | predicted 1 | |
| FrontierSWE | est. | 5% | predicted 0 | |
| CursorBench | est. | 4% | predicted 8 | |
| GSO-Bench | est. | 4% | predicted 0 | |
| METR Time Horizons | est. | 10% | predicted 28 | |
| Vending-Bench 2 | est. | 6% | predicted 0 | |
| PostTrainBench | est. | 4% | predicted 5 | |
| APEX-Agents | est. | 8% | predicted 2 | |
| GDPval | est. | 7% | predicted 5 | |
| Remote Labor Index | est. | 5% | predicted 0 | |
| DeepResearch Bench | est. | 5% | predicted 35 | |
| OSWorld 2.0 | est. | 6% | predicted 0 | |
| OSWorld | est. | 4% | predicted 2 | |
| Cybench | est. | 5% | predicted 3 | |
| CL-bench | est. | 5% | predicted 7 |
Other benchmarks
| Benchmark | Category | Score | Source |
|---|---|---|---|
| DTBench | Knowledge | 50.1% | source |
| GPQA Diamond | Reasoning | 36.3% | source |
| LMCA | Knowledge | 8.8% | source |
| MATH Level 5 | Math | 14.9% | source |
| MMLU | Legacy | 73.8% | source |
| OTIS Mock AIME | Math | 1.8% | source |