Gemini 3.8 Flash
- Input price
- $0.75
- Per 1 million tokens (USD)
- Output price
- $3.75
- Per 1 million tokens (USD)
- Effort levels
- 3total
- Default: Med
- Benchmark results
- 22total
- Benchmark types: 22
Model information
- API model ID
- gemini-3.8-flash
- release date
- 2026-09-02
- effort
- Low / Med (default) / High
- Effort source
- Based on thinking level in Gemini API thinking documentation (aistudio.google.com/docs/thinking).
- Claude Code
- unavailable
- Codex
- unavailable
- Availability notes
- Support in Claude Code / Codex is not explicitly stated officially (Gemini API, AI Studio, Antigravity, etc. are official routes). The official model list for Claude Code / Codex (checked 2026-11-12) contains only Anthropic / OpenAI models, respectively, so both are false.
- Availability source
- https://aistudio.google.com/docs/models
Price (USD per 1 million tokens)
Prices do not vary by effort; token consumption does.
| input | $0.75 |
|---|---|
| output | $3.75 |
| cache read | $0.075 |
| cache write | — |
Introductory price. After revision: input $1.50 / output $7.50 / cache $0.15. Cache storage changes from $0.50 to $1.00 /1M tokens/hr
Gemini 3.8 Flash gemini-3.8-flash $0.75 → $1.50 / $3.75 → $7.50 / $0.075 → $0.15Benchmark results
| Benchmark | effort | Score | Source cost (metric) | measured/reported by | source | Conditions and extraction method |
|---|---|---|---|---|---|---|
| Terminal-Bench 4.0not stated | not stated | 19.1% | — | Gemini 3.8 Flash model cardofficial · Retrieved 2026-11-12 | Vendor-reported value. Self-reported by Google (model card, as of 2026-09). thinking setting not statedtranscribed from tableSource excerptTerminal-bench 4.0 | 19.1% | |
| Terminal-Bench 2.1not stated | not stated | 89.4% | — | Gemini 3.8 Flash model cardofficial · Retrieved 2026-11-12 | Vendor-reported value. Self-reported by Google (model card, as of 2026-09). thinking setting not statedtranscribed from tableSource excerptTerminal-bench 2.1 | 89.4% | |
| DeepSWE v1.1not stated | not stated | 73.7% | — | Gemini 3.8 Flash model cardofficial · Retrieved 2026-11-12 | Vendor-reported value. Self-reported by Google (model card, as of 2026-09). thinking setting not statedtranscribed from tableSource excerptDeepSWE v1.1 | 73.7% | |
| GDPval-AA v2not stated | not stated | 1,545 | — | Gemini 3.8 Flash model cardofficial · Retrieved 2026-11-12 | Vendor-reported value. Self-reported by Google (model card, as of 2026-09). thinking setting not statedtranscribed from tableSource excerptGDPVal-AA v2 (Elo) | 1545 | |
| HLE-Verifiednot stated | not stated | 54.9% | — | Gemini 3.8 Flash model cardofficial · Retrieved 2026-11-12 | Vendor-reported value. Self-reported by Google (model card, as of 2026-09). thinking setting not statedtranscribed from tableSource excerptHLE-Verified | 54.9% | |
| OSWorld 2.0not stated | not stated | 59.0% | — | Gemini 3.8 Flash model cardofficial · Retrieved 2026-11-12 | Vendor-reported value. Self-reported by Google (model card, as of 2026-09). thinking setting not statedtranscribed from tableSource excerptOSWorld-2.0 (partial score, batch tool enabled) | 59.0% | |
| Vals Finance Agent v2not stated | not stated | 61.4% | — | Gemini 3.8 Flash model cardofficial · Retrieved 2026-11-12 | Vendor-reported value. Self-reported by Google (model card, as of 2026-09). thinking setting not statedtranscribed from tableSource excerptVals Finance Agent v2 | 61.4% | |
| Harvey Legal Agent Benchmarknot stated | not stated | 10.0% | — | Gemini 3.8 Flash model cardofficial · Retrieved 2026-11-12 | Vendor-reported value. Self-reported by Google (model card, as of 2026-09). thinking setting not statedtranscribed from tableSource excerptHarvey's Legal Agent Benchmark (all pass rate) | 10.0% | |
| LABBench 2not stated | not stated | 86.2% | — | Gemini 3.8 Flash model cardofficial · Retrieved 2026-11-12 | Vendor-reported value. Self-reported by Google (model card, as of 2026-09). thinking setting not statedtranscribed from tableSource excerptLABBench2 | 86.2% | |
| LVBenchnot stated | not stated | 87.8% | — | Gemini 3.8 Flash model cardofficial · Retrieved 2026-11-12 | Vendor-reported value. Self-reported by Google (model card, as of 2026-09). thinking setting not stated. agentic value (static is 87.1%)transcribed from tableSource excerptLVBench (agentic / static) | 87.8% / 87.1% | |
| GDP.PDFnot stated | not stated | 35.0% | — | Gemini 3.8 Flash model cardofficial · Retrieved 2026-11-12 | Vendor-reported value. Self-reported by Google (model card, as of 2026-09). thinking setting not statedtranscribed from tableSource excerptGDP.PDF (all pass rate) | 35.0% | |
| CharXiv Reasoningnot stated | not stated | 86.2% | — | Gemini 3.8 Flash model cardofficial · Retrieved 2026-11-12 | Vendor-reported value. Self-reported by Google (model card, as of 2026-09). thinking setting not statedtranscribed from tableSource excerptCharXiv Reasoning (no tools) | 86.2% | |
| BioMysteryBench(Human Solvable)not stated | not stated | 88.8% | — | Gemini 3.8 Flash model cardofficial · Retrieved 2026-11-12 | Vendor-reported value. Self-reported by Google (model card, as of 2026-09). thinking setting not statedtranscribed from tableSource excerptBioMysteryBench, Human Solvable | 88.8% | |
| BioMysteryBench(Human Difficult)not stated | not stated | 56.5% | — | Gemini 3.8 Flash model cardofficial · Retrieved 2026-11-12 | Vendor-reported value. Self-reported by Google (model card, as of 2026-09). thinking setting not statedtranscribed from tableSource excerptBioMysteryBench, Human Difficult | 56.5% | |
| GPQA DiamondHigh | High | 95.4% | — | Epoch AI | Epoch AI, Capabilities & benchmarking (benchmark_data.zip, gpqa_diamond.csv) CC BY 4.0third-party · Retrieved 2026-11-12 | Run by Epoch AI (third-party measurement). Source label "Gemini 3.8 Flash (high)". effort is indicated by the suffix _high in Epoch's Model version. Data retrieved 2026-11-12.transcribed from tableSource excerptgpqa_diamond.csv: Model version=gemini-3.8-flash_high, Best score (across scorers)=0.9539141414141414, stderr=0.01401136318210652, Started at=2026-09-02T17:15:36.000ZLocation gpqa_diamond.csv, data row 25 (file line 49), id=Br3mjs2GrYcCwX9AMwcihU, Model version=gemini-3.8-flash_high |
| FrontierMath Tiers 1-3(v2)High | High | 68.4% | — | Epoch AI | Epoch AI, Capabilities & benchmarking (benchmark_data.zip, frontiermath_tiers_1_3_v2.csv) CC BY 4.0third-party · Retrieved 2026-11-12 | Run by Epoch AI (third-party measurement). Source label "Gemini 3.8 Flash (high)". effort is indicated by the suffix _high in Epoch's Model version. Data retrieved 2026-11-12.transcribed from tableSource excerptfrontiermath_tiers_1_3_v2.csv: Model version=gemini-3.8-flash_high, Best score (across scorers)=0.6842105263157895, stderr=0.027582557383453006, Started at=2026-09-02T17:15:36.000ZLocation frontiermath_tiers_1_3_v2.csv, data row 21 (file line 34), id=Utgs6DuaQDDdi98ZWH22qL, Model version=gemini-3.8-flash_high |
| FrontierMath Tier 4(v2)High | High | 22.0% | — | Epoch AI | Epoch AI, Capabilities & benchmarking (benchmark_data.zip, frontiermath_tier_4_v2.csv) CC BY 4.0third-party · Retrieved 2026-11-12 | Run by Epoch AI (third-party measurement). Source label "Gemini 3.8 Flash (high)". effort is indicated by the suffix _high in Epoch's Model version. Data retrieved 2026-11-12.transcribed from tableSource excerptfrontiermath_tier_4_v2.csv: Model version=gemini-3.8-flash_high, Best score (across scorers)=0.21951219512195122, stderr=0.06544589202438408, Started at=2026-09-02T17:15:36.000ZLocation frontiermath_tier_4_v2.csv, data row 21 (file line 34), id=Lgy52ZT4YpFrUqqKVu2qUs, Model version=gemini-3.8-flash_high |
| OTIS Mock AIME 2024-2025High | High | 98.9% | — | Epoch AI | Epoch AI, Capabilities & benchmarking (benchmark_data.zip, otis_mock_aime_2024_2025.csv) CC BY 4.0third-party · Retrieved 2026-11-12 | Run by Epoch AI (third-party measurement). Source label "Gemini 3.8 Flash (high)". effort is indicated by the suffix _high in Epoch's Model version. Data retrieved 2026-11-12.transcribed from tableSource excerptotis_mock_aime_2024_2025.csv: Model version=gemini-3.8-flash_high, Best score (across scorers)=0.9888888888888889, stderr=0.008724007409153888, Started at=2026-09-02T17:15:36.000ZLocation otis_mock_aime_2024_2025.csv, data row 28 (file line 52), id=NeYuqgQyWtFWAHEjQYPm5Z, Model version=gemini-3.8-flash_high |
| SimpleQA VerifiedHigh | High | 69.7% | — | Epoch AI | Epoch AI, Capabilities & benchmarking (benchmark_data.zip, simpleqa_verified.csv) CC BY 4.0third-party · Retrieved 2026-11-12 | Run by Epoch AI (third-party measurement). Source label "Gemini 3.8 Flash (high)". effort is indicated by the suffix _high in Epoch's Model version. Data retrieved 2026-11-12.transcribed from tableSource excerptsimpleqa_verified.csv: Model version=gemini-3.8-flash_high, Best score (across scorers)=0.697, stderr=0.014539683710535193, Started at=2026-09-02T17:15:36.000ZLocation simpleqa_verified.csv, data row 25 (file line 49), id=gZ4bwozTKtpmZTqnqX3a33, Model version=gemini-3.8-flash_high |
| Chess Puzzles(Epoch AI)High | High | 58.0% | — | Epoch AI | Epoch AI, Capabilities & benchmarking (benchmark_data.zip, chess_puzzles.csv) CC BY 4.0third-party · Retrieved 2026-11-12 | Run by Epoch AI (third-party measurement). Source label "Gemini 3.8 Flash (high)". effort is indicated by the suffix _high in Epoch's Model version. Data retrieved 2026-11-12.transcribed from tableSource excerptchess_puzzles.csv: Model version=gemini-3.8-flash_high, Best score (across scorers)=0.58, stderr=0.04960449637488583, Started at=2026-10-13T12:39:05.000ZLocation chess_puzzles.csv, data row 163 (file line 220), id=Hpyy6CpmXmd5q7Zbdq63CR, Model version=gemini-3.8-flash_high |
| MirrorCodeHigh | High | 38.9% | — | Epoch AI | Epoch AI, Capabilities & benchmarking (benchmark_data.zip, mirrorcode.csv) CC BY 4.0third-party · Retrieved 2026-11-12 | Run by Epoch AI (third-party measurement). Source label "Gemini 3.8 Flash (high)". effort is indicated by the suffix _high in Epoch's Model version. Data retrieved 2026-11-12.transcribed from tableSource excerptmirrorcode.csv: Model version=gemini-3.8-flash_high, Best score (across scorers)=0.3888888888888889, stderr=0.11915339216404011, Started at=2026-09-02T17:23:38.000ZLocation mirrorcode.csv, data row 38 (file line 58), id=Fi32v8c7rD82qhgYVcsz7g, Model version=gemini-3.8-flash_high |
| Epoch Capabilities Index(ECI)not stated | not stated | 158.79 | — | Epoch AI | Epoch AI, epoch_capabilities_index/eci_scores.csv (CC BY 4.0)third-party · Retrieved 2026-11-12 | A per-model metric that combines results across multiple benchmarks and multiple efforts (not per effort). Calculated by Epoch AI.transcribed from tableSource excerpteci_scores.csv: Model=Gemini 3.8 Flash, eci=158.79, eci_ci_low=156.77, eci_ci_high=160.94Location epoch_capabilities_index/eci_scores.csv, data row 23 (file line 24), Model=Gemini 3.8 Flash |
Related rankings
- Overall ranking (all models, records)
- Terminal-Bench 4.0 ranking by metric (all models, records)
- Terminal-Bench 2.1 ranking by metric (all models, records)
- DeepSWE v1.1 ranking by metric (all models, records)
- GDPval-AA v2 ranking by metric (all models, records)
- HLE-Verified ranking by metric (all models, records)
- OSWorld 2.0 ranking by metric (all models, records)
- Vals Finance Agent v2 ranking by metric (all models, records)
- Harvey Legal Agent Benchmark ranking by metric (all models, records)
- LABBench 2 ranking by metric (all models, records)
- LVBench ranking by metric (all models, records)
- GDP.PDF ranking by metric (all models, records)
- CharXiv Reasoning ranking by metric (all models, records)
- BioMysteryBench(Human Solvable) ranking by metric (all models, records)
- BioMysteryBench(Human Difficult) ranking by metric (all models, records)
- GPQA Diamond ranking by metric (all models, records)
- FrontierMath Tiers 1-3(v2) ranking by metric (all models, records)
- FrontierMath Tier 4(v2) ranking by metric (all models, records)
- OTIS Mock AIME 2024-2025 ranking by metric (all models, records)
- SimpleQA Verified ranking by metric (all models, records)
- Chess Puzzles(Epoch AI) ranking by metric (all models, records)
- MirrorCode ranking by metric (all models, records)
- Epoch Capabilities Index(ECI) ranking by metric (all models, records)
- Recommendations by role