Language modelActive

Gemini 3 Flash

Google

Released: December 2025
Data date: September 8, 2026

Intelligence: Not ranked
Overall score
Speed: 214.7 tokens/s
Output throughput
Price: $0.5 · $3
Input · Output / 1M tokens
Input: Text, Image, Audio, Video
Output: Text

No published category-wide ranking is available for this model.

Entries 1-12 of 158. Page 1 of 14.

Published benchmarks

Each chart compares models in the same documented test and highlights the current model.

Cybersecurity

2 measurements

MT-AgentRisk

Attack success rate

What this benchmark measures: 365 harmful sequences test safety risks in multi-turn tool use.

Context: A lower attack success rate is better. Capability and safety are separate dimensions.

Lower is better3 modelsGemini 3 Flash: Rank 2 of 3Data date: August 9, 2026

1.Claude Sonnet 4.571.78%

Note: Table 4 reports a 15.43% refusal rate, while Table 9 in the same paper reports 15.34%. This row uses the direct single-turn versus multi-turn comparison table.

2.Gemini 3 Flash78.9%

3.DeepSeek-V3.284.93%

Scale: 0% - 100%

Source: MT-AgentRisk

CC BY 4.0

gradually.ai

Methodology and context

Comparison cohort: mt-agent-risk:multi-turn:asr
Metric: attack success rate
Scale: 0-100
Participants: 3
Methodology and context: OpenHands without a defense layer, judged by GPT-4.1. Single-turn and multi-turn cohorts remain separate.. Source model ID: curated/gemini-3-flash. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: MT-AgentRisk. Dataset type: public. Leaderboard date: 2026-02-13. Harness: OpenHands. Source ID: mt-agent-risk. Evidence capture ID: curated-mt-agent-risk-2026-08-09-1028329f9a3b. Source payload hash: 1028329f9a3b8d4820f4ee86b731973dd37ffa0e4246efe61c195459d652b47f. Parser version: curated-v1. Imported at: 2026-09-04

Source: MT-AgentRisk

MT-AgentRisk

Refusal rate

What this benchmark measures: 365 harmful sequences test safety risks in multi-turn tool use.

Context: A lower attack success rate is better. Capability and safety are separate dimensions.

Higher is better3 modelsGemini 3 Flash: Rank 2 of 3Data date: August 9, 2026

1.Claude Sonnet 4.515.43%

Note: Table 4 reports a 15.43% refusal rate, while Table 9 in the same paper reports 15.34%. This row uses the direct single-turn versus multi-turn comparison table.

2.Gemini 3 Flash6.03%

3.DeepSeek-V3.21.1%

Scale: 0% - 100%

Source: MT-AgentRisk

CC BY 4.0

gradually.ai

Methodology and context

Comparison cohort: mt-agent-risk:multi-turn:refusal
Metric: refusal rate
Scale: 0-100
Participants: 3
Methodology and context: OpenHands without a defense layer, judged by GPT-4.1. Single-turn and multi-turn cohorts remain separate.. Source model ID: curated/gemini-3-flash. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: MT-AgentRisk. Dataset type: public. Leaderboard date: 2026-02-13. Harness: OpenHands. Source ID: mt-agent-risk. Evidence capture ID: curated-mt-agent-risk-2026-08-09-dd3b683c9326. Source payload hash: dd3b683c9326b8ebb99ee225badbd3d228bed7961fbc39502881384ab4cb5667. Parser version: curated-v1. Imported at: 2026-09-04

Source: MT-AgentRisk

Coding

4 measurements

Code Migration

CLI

Higher is better6 modelsGemini 3 Flash: Rank 4 of 6Data date: August 11, 2026

1.Gemini 3.6 Flash19.33%

2.Gemini 3.5 Flash17.18%

3.Gemini 3.1 Pro Preview10.83%

4.Gemini 3 Flash8.52%

5.Gemini 3.5 Flash-Lite5.88%

6.Gemini 3.1 Flash-Lite5.73%

Scale: 0% - 100%

Source: Vals AI

CC BY 4.0

gradually.ai

Methodology and context

Comparison cohort: vals-code-migration:1:cli
Metric: accuracy
Scale: 0-100
Participants: 6
Methodology and context: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-09. Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536. Latency in seconds: 294.005. Standard error: 1.458. Cost per task in USD: 0.353141. Source ID: vals-ai. Evidence capture ID: vals-code-migration-2026-08-11-3a2d5ce718a3. Source payload hash: 3a2d5ce718a34b36c00a1447644df407304422f4ce9a7feb98488c4a04ea39cd. Parser version: vals-astro-v2. Imported at: 2026-08-11

Source: Vals AI

Code Migration

COBOL

Higher is better6 modelsGemini 3 Flash: Rank 6 of 6Data date: August 11, 2026

1.Gemini 3.6 Flash65.71%

2.Gemini 3.5 Flash55.46%

3.Gemini 3.1 Pro Preview36.75%

4.Gemini 3.5 Flash-Lite6.96%

5.Gemini 3.1 Flash-Lite1.25%

6.Gemini 3 Flash0%

Scale: 0% - 100%

Source: Vals AI

CC BY 4.0

gradually.ai

Methodology and context

Comparison cohort: vals-code-migration:1:cobol
Metric: accuracy
Scale: 0-100
Participants: 6
Methodology and context: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-09. Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536. Latency in seconds: 254.923. Standard error: 0. Cost per task in USD: 0.300925. Source ID: vals-ai. Evidence capture ID: vals-code-migration-2026-08-11-3a2d5ce718a3. Source payload hash: 3a2d5ce718a34b36c00a1447644df407304422f4ce9a7feb98488c4a04ea39cd. Parser version: vals-astro-v2. Imported at: 2026-08-11

Source: Vals AI

Code Migration

Code Quality

Higher is better6 modelsGemini 3 Flash: Rank 4 of 6Data date: August 11, 2026

1.Gemini 3.6 Flash3.3 rating

2.Gemini 3.5 Flash2.79 rating

3.Gemini 3.1 Flash-Lite2.66 rating

4.Gemini 3 Flash2.66 rating

5.Gemini 3.5 Flash-Lite2.65 rating

6.Gemini 3.1 Pro Preview2.54 rating

Scale: 0 rating - 3.3 rating

Source: Vals AI

CC BY 4.0

gradually.ai

Methodology and context

Comparison cohort: vals-code-migration:1:code_quality
Metric: mean reviewer code-quality rating
Scale: No upper bound is documented. The chart spans the recorded values and zero.
Participants: 6
Methodology and context: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-09. Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536. Latency in seconds: 284.234. Standard error: 0.047. Cost per task in USD: 0.340087. Score display: 2.662 / 5. Source ID: vals-ai. Evidence capture ID: vals-code-migration-2026-08-11-3a2d5ce718a3. Source payload hash: 3a2d5ce718a34b36c00a1447644df407304422f4ce9a7feb98488c4a04ea39cd. Parser version: vals-astro-v2. Imported at: 2026-08-11

Source: Vals AI

Code Migration

Overall

Higher is better35 modelsGemini 3 Flash: Rank 33 of 35Data date: September 4, 2026

1.GPT-6 Astra67.74%

2.Claude Opus 557.47%

3.Claude Fable 555.07%

4.GPT-5.6 Sol52.92%

5.GPT-5.6 Terra47.8%

6.GPT-5.545.16%

7.Grok 4.644.57%

8.GPT-5.6 Luna44.55%

9.Claude Sonnet 544.39%

10.GLM-5.344.22%

11.GLM-5.237.87%

12.Grok 4.536.6%

13.Gemini 3.8 Flash36.55%

14.GPT-5.434.98%

15.Gemini 3.7 Flash34.8%

16.Gemini 3.6 Flash30.93%

17.Kimi K2.627.77%

18.Gemini 3.5 Flash26.75%

19.DeepSeek-V4-Pro26.2%

20.GLM-5.125.77%

21.Kimi K2.7 Code25.39%

22.MiMo-V2.5-Pro21.56%

23.GLM-5.3-Flash20.52%

24.Gemini 3.1 Pro Preview17.31%

25.Kimi K316.1%

26.GPT-5.4 nano14.47%

27.MiMo-V2.514.23%

28.Inkling Small13.68%

29.GPT-5.4 mini12.94%

30.Inkling11.79%

31.Kimi K2.56.96%

32.Grok 4.36.8%

33.Gemini 3 Flash6.39%

34.Gemini 3.5 Flash-Lite6.15%

35.Gemini 3.1 Flash-Lite4.61%

Scale: 0% - 100%

Source: Vals AI

CC BY 4.0

gradually.ai

Methodology and context

Comparison cohort: vals-code-migration:1:overall
Metric: accuracy
Scale: 0-100
Participants: 35
Methodology and context: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-09-03. Latency in seconds: 284.234. Cost per task in USD: 0.340087. Source ID: vals-ai. Evidence capture ID: vals-code-migration-2026-09-04-37e66450d0d6. Source payload hash: 37e66450d0d6f681e77c4fe9942f74c0aa743078002f50aef50932407223f88f. Parser version: vals-astro-v2. Imported at: 2026-09-04

Source: Vals AI

Finance

6 measurements

CorpFin v2

Exact Pages

Higher is better6 modelsGemini 3 Flash: Rank 1 of 6Data date: August 11, 2026

1.Gemini 3 Flash68.07%

2.Gemini 3.5 Flash62.82%

3.Gemini 3.6 Flash61.07%

4.Gemini 3.1 Pro Preview60.84%

5.Gemini 3.5 Flash-Lite57.46%

6.Gemini 3.1 Flash-Lite56.64%

Scale: 0% - 100%

Source: Vals AI

CC BY 4.0

gradually.ai

Methodology and context

Comparison cohort: vals-corp-fin-v2:2:exact_pages
Metric: accuracy
Scale: 0-100
Participants: 6
Methodology and context: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-06. Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536. Latency in seconds: 11.684. Standard error: 1.592. Cost per task in USD: 0.098516. Source ID: vals-ai. Evidence capture ID: vals-corp_fin_v2-2026-08-11-d9e58d191cd9. Source payload hash: d9e58d191cd9f1690e3b59f0a37055813038531410f2d73e765676893c0f9526. Parser version: vals-astro-v2. Imported at: 2026-08-11

Source: Vals AI

CorpFin v2

Max Fitting Context

Higher is better6 modelsGemini 3 Flash: Rank 4 of 6Data date: August 11, 2026

1.Gemini 3.1 Pro Preview66.9%

2.Gemini 3.5 Flash66.67%

3.Gemini 3.6 Flash65.39%

4.Gemini 3 Flash63.52%

5.Gemini 3.5 Flash-Lite62.35%

6.Gemini 3.1 Flash-Lite60.61%

Scale: 0% - 100%

Source: Vals AI

CC BY 4.0

gradually.ai

Methodology and context

Comparison cohort: vals-corp-fin-v2:2:max_fitting_context
Metric: accuracy
Scale: 0-100
Participants: 6
Methodology and context: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-06. Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536. Latency in seconds: 7.32. Standard error: 1.643. Cost per task in USD: 0.003928. Source ID: vals-ai. Evidence capture ID: vals-corp_fin_v2-2026-08-11-d9e58d191cd9. Source payload hash: d9e58d191cd9f1690e3b59f0a37055813038531410f2d73e765676893c0f9526. Parser version: vals-astro-v2. Imported at: 2026-08-11

Source: Vals AI

CorpFin v2

Overall

Higher is better50 modelsGemini 3 Flash: Rank 12 of 50Data date: September 4, 2026

1.Claude Opus 573.19%

2.Claude Fable 571.83%

3.Kimi K371.56%

4.Inkling Small69.62%

5.Inkling68.57%

6.Grok 4.368.53%

7.GPT-5.568.42%

8.Kimi K2.568.26%

9.Grok 4.567.41%

10.Claude Sonnet 566.98%

11.Kimi K2.666.74%

12.Gemini 3 Flash66.43%

13.Grok 4.666.16%

14.GLM-5.266.12%

15.GPT-5.6 Terra65.31%

16.GPT-5.465.27%

17.Gemini 3.5 Flash64.69%

18.Gemini 3.1 Pro Preview64.49%

19.GLM-5.164.45%

20.GPT-5.6 Sol64.38%

21.GPT-5.6 Luna64.22%

22.Gemini 3.6 Flash63.33%

23.GPT-4.163.05%

24.GLM-562.9%

25.MiMo-V2.5-Pro61.42%

26.DeepSeek-V4-Pro61.38%

27.GPT-5.4 nano61.19%

28.GPT-561.07%

29.GLM-4.560.96%

30.GPT-5.4 mini60.92%

31.Gemini 2.5 Pro60.8%

32.Gemini 3.5 Flash-Lite60.68%

33.Kimi K2 Thinking60.57%

34.GPT-5 mini60.18%

35.MiMo-V2.559.91%

36.o359.71%

37.Gemini 3.1 Flash-Lite59.36%

38.o4-mini58.97%

39.GPT OSS 120B58.24%

40.GPT-4.1 mini57.93%

41.GLM-4.656.84%

42.DeepSeek-V3 032454.74%

43.Trinity Large Thinking54.66%

44.DeepSeek-R154.12%

45.GPT OSS 20B53.15%

46.DeepSeek-V352.49%

47.GLM-4.746.39%

48.GPT-4o mini45.46%

49.o3-mini45.3%

50.GPT-4.1 nano42.08%

Scale: 0% - 100%

Source: Vals AI

CC BY 4.0

gradually.ai

Methodology and context

Comparison cohort: vals-corp-fin-v2:2:overall
Metric: accuracy
Scale: 0-100
Participants: 50
Methodology and context: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-12. Latency in seconds: 11.279. Cost per task in USD: 0.047154. Source ID: vals-ai. Evidence capture ID: vals-corp_fin_v2-2026-09-04-46dad6684551. Source payload hash: 46dad668455190ee4239b7c743d00d81c6fd3127e0d87123c518ec107c13b44d. Parser version: vals-astro-v2. Imported at: 2026-09-04

Source: Vals AI

CorpFin v2

Shared Max Context

Higher is better6 modelsGemini 3 Flash: Rank 1 of 6Data date: August 11, 2026

1.Gemini 3 Flash67.72%

2.Gemini 3.1 Pro Preview65.73%

3.Gemini 3.5 Flash64.57%

4.Gemini 3.6 Flash63.52%

5.Gemini 3.5 Flash-Lite62.24%

6.Gemini 3.1 Flash-Lite60.84%

Scale: 0% - 100%

Source: Vals AI

CC BY 4.0

gradually.ai

Methodology and context

Comparison cohort: vals-corp-fin-v2:2:shared_max_context
Metric: accuracy
Scale: 0-100
Participants: 6
Methodology and context: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-06. Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536. Latency in seconds: 14.833. Standard error: 1.596. Cost per task in USD: 0.039017. Source ID: vals-ai. Evidence capture ID: vals-corp_fin_v2-2026-08-11-d9e58d191cd9. Source payload hash: d9e58d191cd9f1690e3b59f0a37055813038531410f2d73e765676893c0f9526. Parser version: vals-astro-v2. Imported at: 2026-08-11

Source: Vals AI

EMB

Overall

Higher is better32 modelsGemini 3 Flash: Rank 30 of 32Data date: September 4, 2026

1.Claude Fable 573.67%

2.Claude Opus 573.56%

3.GPT-5.6 Sol72.34%

4.Gemini 3.8 Flash72.2%

5.GPT-6 Astra71.7%

6.Gemini 3.7 Flash71.33%

7.GPT-5.6 Luna67.12%

8.Kimi K366.4%

9.Claude Sonnet 566.32%

10.GPT-5.6 Terra66.21%

11.Gemini 3.6 Flash65.41%

12.GPT-5.564.54%

13.Gemini 3.5 Flash63.55%

14.Grok 4.662.73%

15.GLM-5.261.53%

16.Kimi K2.657.86%

17.GLM-5.356.34%

18.GLM-5.3-Flash55.93%

19.MiMo-V2.5-Pro55.23%

20.MiMo-V2.555.09%

21.Grok 4.552.91%

22.Gemini 3.1 Pro Preview52.62%

23.DeepSeek-V4-Pro51.62%

24.GPT-5.4 mini45.43%

25.GPT-5.4 nano44.75%

26.Gemini 3.5 Flash-Lite43.23%

27.Inkling38.36%

28.Inkling Small31.76%

29.Kimi K2.528.47%

30.Gemini 3 Flash26.02%

31.Grok 4.318.39%

32.Gemini 3.1 Flash-Lite8.63%

Scale: 0% - 100%

Source: Vals AI

CC BY 4.0

gradually.ai

Methodology and context

Comparison cohort: vals-emb:1:overall
Metric: accuracy
Scale: 0-100
Participants: 32
Methodology and context: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-09-03. Latency in seconds: 328.69. Cost per task in USD: 0.453183. Source ID: vals-ai. Evidence capture ID: vals-emb-2026-09-04-5d630bb9b3a7. Source payload hash: 5d630bb9b3a7674ec91efa385ebf472ae6d8bb96c8db2fa2dd5af3e813e66155. Parser version: vals-astro-v2. Imported at: 2026-09-04

Source: Vals AI

Finance Agent (v2)

Adjustments

Higher is better6 modelsGemini 3 Flash: Rank 5 of 6Data date: August 11, 2026

1.Gemini 3.5 Flash54.87%

2.Gemini 3.6 Flash53.61%

3.Gemini 3.5 Flash-Lite41.74%

4.Gemini 3.1 Pro Preview38.36%

5.Gemini 3 Flash35.93%

6.Gemini 3.1 Flash-Lite30.97%

Scale: 0% - 100%

Source: Vals AI

CC BY 4.0

gradually.ai

Methodology and context

Comparison cohort: vals-finance-agent-v2:2:adjustments
Metric: accuracy
Scale: 0-100
Participants: 6
Methodology and context: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-06. Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536. Latency in seconds: 226.691. Standard error: 1.014. Cost per task in USD: 0.50541. Source ID: vals-ai. Evidence capture ID: vals-fabv2-2026-08-11-3e5c93a1857c. Source payload hash: 3e5c93a1857cb9afa3ea6c689b69ffc95e4399fdbacb84e03b975b7b8383d7c1. Parser version: vals-astro-v2. Imported at: 2026-08-11

Source: Vals AI

Recorded measurement details

Every observation retains its source value and published test conditions.

MT-AgentRisk: Attack success rate (Show measurement, test conditions, and source)

Source value: 78.9
Score: 78.9
Metric: attack success rate
Unit: %
Benchmark version: ICML 2026 paper
Category: cybersecurity
Task: asr
Task label: Attack success rate
Direction: Lower is better
Comparison cohort: mt-agent-risk:multi-turn:asr
Source type: independent-evaluator
Evaluator: MT-AgentRisk
Status: active
Retrieved at: 2026-08-09
Methodology: OpenHands without a defense layer, judged by GPT-4.1. Single-turn and multi-turn cohorts remain separate.. Source model ID: curated/gemini-3-flash. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: MT-AgentRisk. Dataset type: public. Leaderboard date: 2026-02-13. Harness: OpenHands. Source ID: mt-agent-risk. Evidence capture ID: curated-mt-agent-risk-2026-08-09-1028329f9a3b. Source payload hash: 1028329f9a3b8d4820f4ee86b731973dd37ffa0e4246efe61c195459d652b47f. Parser version: curated-v1. Imported at: 2026-09-04
Comparison methodology: OpenHands without a defense layer, judged by GPT-4.1. Single-turn and multi-turn cohorts remain separate.

Source MT-AgentRisk

MT-AgentRisk: Refusal rate (Show measurement, test conditions, and source)

Source value: 6.03
Score: 6.03
Metric: refusal rate
Unit: %
Benchmark version: ICML 2026 paper
Category: cybersecurity
Task: refusal
Task label: Refusal rate
Direction: Higher is better
Comparison cohort: mt-agent-risk:multi-turn:refusal
Source type: independent-evaluator
Evaluator: MT-AgentRisk
Status: active
Retrieved at: 2026-08-09
Methodology: OpenHands without a defense layer, judged by GPT-4.1. Single-turn and multi-turn cohorts remain separate.. Source model ID: curated/gemini-3-flash. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: MT-AgentRisk. Dataset type: public. Leaderboard date: 2026-02-13. Harness: OpenHands. Source ID: mt-agent-risk. Evidence capture ID: curated-mt-agent-risk-2026-08-09-dd3b683c9326. Source payload hash: dd3b683c9326b8ebb99ee225badbd3d228bed7961fbc39502881384ab4cb5667. Parser version: curated-v1. Imported at: 2026-09-04
Comparison methodology: OpenHands without a defense layer, judged by GPT-4.1. Single-turn and multi-turn cohorts remain separate.

Source MT-AgentRisk

Code Migration: CLI (Show measurement, test conditions, and source)

Source value: 8.521
Score: 8.521
Metric: accuracy
Unit: %
Benchmark version: 1
Category: coding
Task: cli
Task label: CLI
Direction: Higher is better
Comparison cohort: vals-code-migration:1:cli
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-08-11
Methodology: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-09. Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536. Latency in seconds: 294.005. Standard error: 1.458. Cost per task in USD: 0.353141. Source ID: vals-ai. Evidence capture ID: vals-code-migration-2026-08-11-3a2d5ce718a3. Source payload hash: 3a2d5ce718a34b36c00a1447644df407304422f4ce9a7feb98488c4a04ea39cd. Parser version: vals-astro-v2. Imported at: 2026-08-11
Comparison methodology: Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536

Source Vals AI

Code Migration: COBOL (Show measurement, test conditions, and source)

Source value: 0
Score: 0
Metric: accuracy
Unit: %
Benchmark version: 1
Category: coding
Task: cobol
Task label: COBOL
Direction: Higher is better
Comparison cohort: vals-code-migration:1:cobol
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-08-11
Methodology: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-09. Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536. Latency in seconds: 254.923. Standard error: 0. Cost per task in USD: 0.300925. Source ID: vals-ai. Evidence capture ID: vals-code-migration-2026-08-11-3a2d5ce718a3. Source payload hash: 3a2d5ce718a34b36c00a1447644df407304422f4ce9a7feb98488c4a04ea39cd. Parser version: vals-astro-v2. Imported at: 2026-08-11
Comparison methodology: Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536

Source Vals AI

Code Migration: Code Quality (Show measurement, test conditions, and source)

Source value: 2.662
Score: 2.662
Metric: mean reviewer code-quality rating
Unit: rating
Benchmark version: 1
Category: coding
Task: code_quality
Task label: Code Quality
Direction: Higher is better
Comparison cohort: vals-code-migration:1:code_quality
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-08-11
Methodology: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-09. Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536. Latency in seconds: 284.234. Standard error: 0.047. Cost per task in USD: 0.340087. Score display: 2.662 / 5. Source ID: vals-ai. Evidence capture ID: vals-code-migration-2026-08-11-3a2d5ce718a3. Source payload hash: 3a2d5ce718a34b36c00a1447644df407304422f4ce9a7feb98488c4a04ea39cd. Parser version: vals-astro-v2. Imported at: 2026-08-11
Comparison methodology: Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536

Source Vals AI

Code Migration: Overall (Show measurement, test conditions, and source)

Source value: 6.391
Score: 6.391
Metric: accuracy
Unit: %
Benchmark version: 1
Category: coding
Task: overall
Task label: Overall
Direction: Higher is better
Comparison cohort: vals-code-migration:1:overall
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-09-04
Methodology: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-09-03. Latency in seconds: 284.234. Cost per task in USD: 0.340087. Source ID: vals-ai. Evidence capture ID: vals-code-migration-2026-09-04-37e66450d0d6. Source payload hash: 37e66450d0d6f681e77c4fe9942f74c0aa743078002f50aef50932407223f88f. Parser version: vals-astro-v2. Imported at: 2026-09-04

Source Vals AI

CorpFin v2: Exact Pages (Show measurement, test conditions, and source)

Source value: 68.065
Score: 68.065
Metric: accuracy
Unit: %
Benchmark version: 2
Category: finance
Task: exact_pages
Task label: Exact Pages
Direction: Higher is better
Comparison cohort: vals-corp-fin-v2:2:exact_pages
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-08-11
Methodology: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-06. Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536. Latency in seconds: 11.684. Standard error: 1.592. Cost per task in USD: 0.098516. Source ID: vals-ai. Evidence capture ID: vals-corp_fin_v2-2026-08-11-d9e58d191cd9. Source payload hash: d9e58d191cd9f1690e3b59f0a37055813038531410f2d73e765676893c0f9526. Parser version: vals-astro-v2. Imported at: 2026-08-11
Comparison methodology: Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536

Source Vals AI

CorpFin v2: Max Fitting Context (Show measurement, test conditions, and source)

Source value: 63.52
Score: 63.52
Metric: accuracy
Unit: %
Benchmark version: 2
Category: finance
Task: max_fitting_context
Task label: Max Fitting Context
Direction: Higher is better
Comparison cohort: vals-corp-fin-v2:2:max_fitting_context
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-08-11
Methodology: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-06. Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536. Latency in seconds: 7.32. Standard error: 1.643. Cost per task in USD: 0.003928. Source ID: vals-ai. Evidence capture ID: vals-corp_fin_v2-2026-08-11-d9e58d191cd9. Source payload hash: d9e58d191cd9f1690e3b59f0a37055813038531410f2d73e765676893c0f9526. Parser version: vals-astro-v2. Imported at: 2026-08-11
Comparison methodology: Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536

Source Vals AI

CorpFin v2: Overall (Show measurement, test conditions, and source)

Source value: 66.434
Score: 66.434
Metric: accuracy
Unit: %
Benchmark version: 2
Category: finance
Task: overall
Task label: Overall
Direction: Higher is better
Comparison cohort: vals-corp-fin-v2:2:overall
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-09-04
Methodology: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-12. Latency in seconds: 11.279. Cost per task in USD: 0.047154. Source ID: vals-ai. Evidence capture ID: vals-corp_fin_v2-2026-09-04-46dad6684551. Source payload hash: 46dad668455190ee4239b7c743d00d81c6fd3127e0d87123c518ec107c13b44d. Parser version: vals-astro-v2. Imported at: 2026-09-04

Source Vals AI

CorpFin v2: Shared Max Context (Show measurement, test conditions, and source)

Source value: 67.716
Score: 67.716
Metric: accuracy
Unit: %
Benchmark version: 2
Category: finance
Task: shared_max_context
Task label: Shared Max Context
Direction: Higher is better
Comparison cohort: vals-corp-fin-v2:2:shared_max_context
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-08-11
Methodology: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-06. Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536. Latency in seconds: 14.833. Standard error: 1.596. Cost per task in USD: 0.039017. Source ID: vals-ai. Evidence capture ID: vals-corp_fin_v2-2026-08-11-d9e58d191cd9. Source payload hash: d9e58d191cd9f1690e3b59f0a37055813038531410f2d73e765676893c0f9526. Parser version: vals-astro-v2. Imported at: 2026-08-11
Comparison methodology: Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536

Source Vals AI

EMB: Overall (Show measurement, test conditions, and source)

Source value: 26.017
Score: 26.017
Metric: accuracy
Unit: %
Benchmark version: 1
Category: finance
Task: overall
Task label: Overall
Direction: Higher is better
Comparison cohort: vals-emb:1:overall
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-09-04
Methodology: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-09-03. Latency in seconds: 328.69. Cost per task in USD: 0.453183. Source ID: vals-ai. Evidence capture ID: vals-emb-2026-09-04-5d630bb9b3a7. Source payload hash: 5d630bb9b3a7674ec91efa385ebf472ae6d8bb96c8db2fa2dd5af3e813e66155. Parser version: vals-astro-v2. Imported at: 2026-09-04

Source Vals AI

Finance Agent (v2): Adjustments (Show measurement, test conditions, and source)

Source value: 35.931
Score: 35.931
Metric: accuracy
Unit: %
Benchmark version: 2
Category: finance
Task: adjustments
Task label: Adjustments
Direction: Higher is better
Comparison cohort: vals-finance-agent-v2:2:adjustments
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-08-11
Methodology: Source model ID: google/gemini-3-flash-preview. Source model name: Gemini 3 Flash. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-06. Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536. Latency in seconds: 226.691. Standard error: 1.014. Cost per task in USD: 0.50541. Source ID: vals-ai. Evidence capture ID: vals-fabv2-2026-08-11-3e5c93a1857c. Source payload hash: 3e5c93a1857cb9afa3ea6c689b69ffc95e4399fdbacb84e03b975b7b8383d7c1. Parser version: vals-astro-v2. Imported at: 2026-08-11
Comparison methodology: Reasoning effort: high. Temperature: 1. Maximum output tokens: 65536

Source Vals AI

Our model tests

Community tests you can inspect

Compare how the models respond to the same prompt. Each test shows the first attempt, with no subsequent fixes to the generated code. These results do not contribute to an overall score.

Pelican on a bicycle

A community classic for free-form SVG drawing.

Gemini 3 Flash

Not tested yet

Prompt and test conditions

Generate an SVG of a pelican riding a bicycle

Token limit including reasoning: 8,192 tokens

Task origin (Simon Willison)

Evidence

Sources and data date

Every statement links to its underlying documentation or leaderboard.

Speed: Google, Non-reasoning. Output speed after the first streaming chunk, not total response time. Classified as deprecated by Artificial Analysis. Artificial Analysis maintains only the default 10K-input workload there. Retrieved September 8, 2026.
Artificial Analysis Methodology
Price: Base API prices without caching or batch discounts. Higher context tiers and other rates are listed under Costs.
Pricing
Input: Documented modalities. Retrieved September 8, 2026.
Model documentation
Output: Documented modalities. Retrieved September 8, 2026.
Model documentation

Research date September 8, 2026. 6 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication.

Open evidence questions and research notes (original wording)

Not found in the inspected sources

parameters-billions. Google does not publish a parameter count on the Gemini API model page.

Not found in the inspected sources

active-parameters-billions. Google does not publish active parameter counts for Gemini API models.

Model metadataSource
Context window data (retrieved September 8, 2026)Source
Knowledge cutoff dataSource
API pricingSource
MT-AgentRisk (retrieved August 9, 2026)Source
Video-MME-v2 (retrieved August 9, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved September 4, 2026)Source
Vals AI (retrieved September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved August 11, 2026; September 4, 2026)Source
Vals AI (retrieved September 4, 2026)Source
Google DeepMind (retrieved September 8, 2026)Source
Artificial Analysis (retrieved September 8, 2026)Source
Model metadata (retrieved September 8, 2026)Source
Speed measurement methodologySource

Gemini 3 Flash

Published benchmarks

Cybersecurity

MT-AgentRisk

MT-AgentRisk

Coding

Code Migration

Code Migration

Code Migration

Code Migration

Finance

CorpFin v2

CorpFin v2

CorpFin v2

CorpFin v2

EMB

Finance Agent (v2)

Recorded measurement details

Community tests you can inspect

Pelican on a bicycle

Specifications and access

Additional source observations

Published prices

100 API requests

Compare this model

Sources and data date