AmazonAuto-detectedNova 2.0 Pro Preview
Best configuration ranks #251 of 610 on the BenchLeader Index at 52.3 ±7.9 (medium reasoning effort). Last measured 2 Dec 2025stale: no new result in six months. Released 2 Dec 2025.
- Blended price
- $3.44/M
- $1.25 in · $10.00 out
- Output speed
- 120 tok/s
- First answer
- 29 s
- first token 1.03 s
- Context
- 256k
- Overall index52
- Coding48
- Agents & tools55
- Knowledge41
- Instruction following76
- Multimodal49
- Long context58
- Composite47
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Default” means the publisher did not say which setting was used.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| no reasoning | – | – | 104 tok/s | 1.03 s | $0.0035 | Coding 28 |
| low | 51.3 | #268 | 120 tok/s | 26 s | $0.0035 | Agents & tools 49 · Coding 42 · Composite 45 · Instruction following 77 · Knowledge 45 · Long context 59 · Multimodal 47 |
| mediumbest | 52.3 | #251 | 120 tok/s | 29 s | $0.0035 | Agents & tools 55 · Coding 48 · Composite 47 · Instruction following 76 · Knowledge 41 · Long context 58 · Multimodal 49 |
| default | 46.7 | #365 | 104 tok/s | 1.03 s | $0.0035 | Agents & tools 48 · Composite 42 · Instruction following 53 · Knowledge 41 · Long context 40 |
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | no reasoning | low | medium | default | Source |
|---|---|---|---|---|---|
| GPQA Diamond (AA)not in index | – | 75.0%#243 | 78.5%#203 | 63.6%#335 | Artificial Analysis |
| Humanity's Last Exam (AA)not in index | – | 5.2%#365 | 9.4%#268 | 3.9%#471 | Artificial Analysis |
Agents & tools
| Benchmark | no reasoning | low | medium | default | Source |
|---|---|---|---|---|---|
| Terminal-Bench Hard | – | 17.4%#173 | 24.2%#137 | 16.7%#179 | Artificial Analysis |
| τ²-Bench Telecom (AA)not in index | – | 90.6%#62 | 92.7%#51 | 71.6%#146 | Artificial Analysis |
Knowledge
| Benchmark | no reasoning | low | medium | default | Source |
|---|---|---|---|---|---|
| AA-Omniscience | – | -40.5#256 | -49.0#310 | -49.5#313 | Artificial Analysis |
Instruction following
| Benchmark | no reasoning | low | medium | default | Source |
|---|---|---|---|---|---|
| IFBench | – | 79.6%#10 | 79.0%#12 | 52.0%#160 | Artificial Analysis |
Multimodal
| Benchmark | no reasoning | low | medium | default | Source |
|---|---|---|---|---|---|
| MMMU-Pro | – | 62.7%#160 | 64.5%#151 | – | Artificial Analysis |
Long context
| Benchmark | no reasoning | low | medium | default | Source |
|---|---|---|---|---|---|
| AA-LCR | – | 64.7%#195 | 64.0%#199 | 30.0%#323 | Artificial Analysis |
Composite
| Benchmark | no reasoning | low | medium | default | Source |
|---|---|---|---|---|---|
| AA Intelligence Index | – | 12.8#256 | 14.2#234 | 10.0#312 | Artificial Analysis |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0035 | – | 31.6 s |
| Summarise a 30-page report | 12,000 / 600 | $0.021 | – | 34.1 s |
| Code edit | 6,000 / 1,500 | $0.022 | – | 41.6 s |
| Agentic coding session | 60,000 / 4,000 | $0.115 | – | 1.0 min |
| Structured extraction | 2,000 / 200 | $0.0045 | – | 30.8 s |
See also
Data as of 9 Sept 2026. Compare these configurations.