GoogleAuto-detectedGemini 2.5 Flash Lite 09
Best configuration ranks #314 of 610 on the BenchLeader Index at 48.8 ±3.1 (thinking reasoning effort). Last measured 1 Sept 2026.
- Blended price
- $0.175/M
- $0.100 in · $0.400 out
- Output speed
- –
- First answer
- –
- Context
- 1M
- Overall index49
- Coding50
- Agents & tools45
- Maths44
- Knowledge47
- Instruction following53
- Multimodal49
- Long context59
- Composite42
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 |
|---|---|---|---|---|---|---|
| thinkingbest | 48.8 | #314 | – | – | $0.0002 | Agents & tools 45 · Coding 50 · Composite 42 · Instruction following 53 · Knowledge 47 · Long context 59 · Maths 44 · Multimodal 49 |
| default | 45.7 | #384 | – | – | $0.0002 | Agents & tools 40 · Coding 47 · Composite 41 · Instruction following 44 · Knowledge 46 · Long context 51 · Maths 42 · Multimodal 48 |
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 | thinking | default | Source |
|---|---|---|---|
| GPQA Diamond (AA)not in index | 70.9%#281 | 65.0%#328 | Artificial Analysis |
| Humanity's Last Exam (AA)not in index | 7.0%#304 | 5.1%#369 | Artificial Analysis |
| GPQA Diamond (Vals)not in index | 70.2%#89 | 64.1%#101 | Vals AI |
Coding
| Benchmark | thinking | default | Source |
|---|---|---|---|
| LiveCodeBench | 71.4%#83 | 67.7%#92 | Vals AI |
Agents & tools
| Benchmark | thinking | default | Source |
|---|---|---|---|
| Terminal-Bench Hard | 12.9%#199 | 7.6%#232 | Artificial Analysis |
| τ²-Bench Telecom (AA)not in index | 30.7%#249 | 30.4%#251 | Artificial Analysis |
Maths
| Benchmark | thinking | default | Source |
|---|---|---|---|
| AIME (Vals) | 42.1%#66 | 26.3%#75 | Vals AI |
| MGSM | 88.4%#50 | 89.5%#43 | Vals AI |
Knowledge
| Benchmark | thinking | default | Source |
|---|---|---|---|
| AA-Omniscience | -54.0#349 | -42.6#269 | Artificial Analysis |
| MMLU-Pro | 79.1%#95 | 78.6%#98 | Vals AI |
| LegalBench | 82.0%#70 | 79.0%#92 | Vals AI |
| CorpFin | 57.6%#85 | 56.3%#88 | Vals AI |
| TaxEval | 64.7%#113 | 66.2%#111 | Vals AI |
| MedQA | 88.9%#51 | 80.3%#70 | Vals AI |
Instruction following
| Benchmark | thinking | default | Source |
|---|---|---|---|
| IFBench | 52.6%#157 | 41.8%#237 | Artificial Analysis |
Multimodal
| Benchmark | thinking | default | Source |
|---|---|---|---|
| MMMU-Pro | 65.0%#145 | 63.4%#156 | Artificial Analysis |
Long context
| Benchmark | thinking | default | Source |
|---|---|---|---|
| AA-LCR | 64.7%#195 | 49.7%#256 | Artificial Analysis |
Composite
| Benchmark | thinking | default | Source |
|---|---|---|---|
| AA Intelligence Index | 10.3#307 | 9.3#332 | 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 | $0.0002 | – | – |
| Summarise a 30-page report | 12,000 / 600 | $0.0014 | – | – |
| Code edit | 6,000 / 1,500 | $0.0012 | – | – |
| Agentic coding session | 60,000 / 4,000 | $0.0076 | – | – |
| Structured extraction | 2,000 / 200 | $0.0003 | – | – |
See also
Data as of 9 Sept 2026. Compare these configurations.