Claude Fable 5 vs Gemini 3 Flash
Full profiles: Claude Fable 5 (Anthropic) and Gemini 3 Flash (Google): every benchmark, price and rank.
Updated · Data from 10 sources
As of 29 Oct 2026, Claude Fable 5 is ahead of Gemini 3 Flash on quality (66.6 vs 59.0 on the BenchLeader Index), but Gemini 3 Flash is 18× cheaper at $1.13 vs $20.00 per 1M tokens blended. Claude Fable 5 leads in every category we measure both in. Gemini 3 Flash streams 3×+ faster.
Key facts (as of 29 Oct 2026)
- Claude Fable 5 scores 66.6 on the BenchLeader Index and Gemini 3 Flash scores 59.0 (BenchLeader, 29 Oct 2026).
- Claude Fable 5 costs $20.00 per 1M tokens blended (list price, 29 Oct 2026).
- Gemini 3 Flash costs $1.13 per 1M tokens blended (list price, 29 Oct 2026).
- Gemini 3 Flash streams 3×+ faster than Claude Fable 5 (Artificial Analysis, 7-day median, 29 Oct 2026).
Claude Fable 5 benchmarks Anthropic
Gemini 3 Flash benchmarks thinking GoogleOpen in the comparison toolWhich to pick for…
The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.
| Use case | Pick | Why | Confidence |
|---|---|---|---|
| Agents & tool use | Claude Fable 5 | 64 vs 55 category score | High |
| Chat quality (human preference) | – | Only Claude Fable 5 has an Arena rating (68 category score) | Not comparable |
| Lowest cost | Gemini 3 Flash | $1.13 vs $20.00 per 1M blended; agentic coding session $0.042 vs $0.800 | – |
| Fastest first answer | Claude Fable 5 | 3.8 s (Claude Fable 5) vs 6.0 s (Gemini 3 Flash (thinking)) to the first answer token | – |
| High volume on a budget | Neither flagship | A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs | – |
Key differences
- Claude Fable 5 leads on quality: 66.6 vs 59.0.
- Claude Fable 5 is stronger in agents & tools, coding, human preference, knowledge, reasoning.
- Gemini 3 Flash (thinking) is stronger in instruction following, long context, multimodal.
- Gemini 3 Flash (thinking) is 18× cheaper ($1.13 vs $20.00 per 1M blended).
- Gemini 3 Flash (thinking) streams 3.9× faster (209 vs 53 tokens per second).
At a glance
| Spec | Claude Fable 5 | Gemini 3 Flash |
|---|---|---|
| Developer | Anthropic | |
| Released | 9 Jun 2026 | 17 Dec 2025 |
| Context window (tokens) | 1,000,000 | 1,048,576 |
| Compared at | Claude Fable 5 | Gemini 3 Flash (thinking) |
| Open weights | No | No |
| Price in / out / cached input per 1M | $10.00 / $50.00 / $1.00 | $0.500 / $3.00 / $0.050 |
Category by category
Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.
| Category | Claude Fable 5 | Gemini 3 Flash (thinking) | Leader |
|---|---|---|---|
| Reasoning | 80.0 | – | not comparable |
| Coding | 70.4 | – | not comparable |
| Agents & tools | 64.2 | 55.1 | Claude Fable 5 |
| Knowledge | 66.5 | – | not comparable |
| Instruction following | – | 75.8 | not comparable |
| Human preference | 67.7 | – | not comparable |
| Multimodal | – | 62.7 | not comparable |
| Long context | – | 63.8 | not comparable |
Where sources disagree
Across the 0 benchmarks both have, Claude Fable 5 leads 0, Gemini 3 Flash leads 0. Every overall index that rates both puts Claude Fable 5 ahead.
| Overall index | Claude Fable 5 | Gemini 3 Flash (thinking) | Leader |
|---|---|---|---|
| BenchLeader Index | 66.6 | 59.0 | Claude Fable 5 |
| Arena Text | 1504† | 1458 | Claude Fable 5 |
Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).
Price, speed and latency
List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page; suite cost is what Artificial Analysis paid to run its whole Intelligence Index.
| Measure | Claude Fable 5 | Gemini 3 Flash (thinking) |
|---|---|---|
| Blended price per 1M tokens | $20.00 | $1.13 |
| Input / output per 1M | $10.00 / $50.00 | $0.500 / $3.00 |
| Cached input per 1M | $1.00 | $0.050 |
| Price per task: chat reply | $0.019 | $0.0011 |
| Price per task: summarize a 30-page report | $0.150 | $0.0078 |
| Price per task: code edit | $0.135 | $0.0075 |
| Price per task: agentic coding session | $0.800 | $0.042 |
| Price per task: structured extraction | $0.030 | $0.0016 |
| Output speed | 53 tok/s | 209 tok/s |
| Time to first answer | 3.8 s | 6.0 s |
| Time to a complete answer | – | 8.4 s |
Each model is compared at its best-ranked reasoning-effort setting. Claude Fable 5 scores 66.6 and takes 4 s to its first answer token; Claude Fable 5 (max) scores 65.3 in 70 s, Claude Fable 5 (high) scores 64.5 in 4 s. Gemini 3 Flash (thinking) scores 59.0 and takes 6 s to its first answer token; Gemini 3 Flash (high) scores 58.3 in 1 s, Gemini 3 Flash scores 57.0 in 1 s. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.
Every benchmark
137 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.
| Metric | Publisher | Claude Fable 5 | Gemini 3 Flash (thinking) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 66.6 | 59.0 |
| Agents & tools score | BenchLeader | 64.2 | 55.1 |
| Coding score | BenchLeader | 70.4 | – |
| Human preference score | BenchLeader | 67.7 | – |
| Knowledge score | BenchLeader | 66.5 | – |
| Reasoning score | BenchLeader | 80.0 | – |
| Instruction following score | BenchLeader | – | 75.8 |
| Long context score | BenchLeader | – | 63.8 |
| Multimodal score | BenchLeader | – | 62.7 |
| Blended price $/M | BenchLeader | $20.00 | $1.13 |
| Output speed | BenchLeader | 53 tok/s | 209 tok/s |
| Time to first answer | BenchLeader | 3.8 s | 6.0 s |
| Context window | BenchLeader | 1,000,000 tokens | 1,048,576 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | 85.9%† | 83.2%† |
| SimpleBench | SimpleBench | 81.9%† | 61.1%† |
| Arena Hard Prompts | Arena (LMArena) | 1529† | 1493† |
| LiveBench 1.0 Reasoning | LiveBench | 89.7%† | – |
| GPQA Diamond (AA) | Artificial Analysis | 92.6%† | 89.8% |
| Humanity's Last Exam (AA) | Artificial Analysis | 55.5%† | 36.6% |
| GPQA Diamond (Vals) | Vals AI | 93.2% | 87.9%† |
| EnigmaEval | Scale AI SEAL | 39.3%† | – |
| Kagi LLM Benchmark | Kagi LLM Benchmark | 88.8% | – |
| ARC-AGI-1 | ARC Prize | 98.5%† | 84.7%† |
| ARC-AGI-2 | ARC Prize | 89.2%† | 33.6%† |
| CritPt | Artificial Analysis | 28.6%† | 8.6% |
| SciPredict | Scale AI SEAL | – | 22.2%† |
| Chess Puzzles | Epoch AI Benchmarking Hub | 47.0%† | 39.0%† |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 52.0%† | 20.0%† |
| LMCA | Epoch AI Benchmarking Hub | 60.3%† | 43.1%† |
| DTBench | Epoch AI Benchmarking Hub | 98.4%† | 89.1%† |
| ForecastBench | ForecastBench | 66.0%† | 64.7%† |
| BTF-3 | Epoch AI Benchmarking Hub | 13.0%† | – |
| Coding | |||
| SWE-bench Verified (Epoch) | Epoch AI Benchmarking Hub | – | 75.4%† |
| SciCode | SciCode | 61.0%† | – |
| WeirdML | WeirdML | 91.9%† | 61.6%† |
| FrontierCode | Cognition | 53.5% | – |
| GSO-Bench | GSO-Bench | 78.4% | 9.8%† |
| Arena Coding | Arena (LMArena) | 1550† | 1505† |
Show all 137 benchmarks
| Metric | Publisher | Claude Fable 5 | Gemini 3 Flash (thinking) |
|---|---|---|---|
| Coding | |||
| Arena WebDev | Arena (LMArena) | 1624† | 1440† |
| LiveBench 1.0 Coding | LiveBench | 86.0%† | – |
| LiveCodeBench | Vals AI | 89.8% | 85.6%† |
| SWE-bench (Vals) | Vals AI | 95.0% | 75.0%† |
| SWE-Bench Pro | Scale AI SEAL | – | 34.6%† |
| SWE-bench Verified (bash only) | SWE-bench | – | 75.8%† |
| SWE-bench Verified (any scaffold) | SWE-bench | – | 75.8%† |
| Code Migration | Vals AI | 55.1% | 6.4%† |
| ProgramBench | Vals AI | 2.0% | 0.0%† |
| Vibe Code Bench v1.1 | Vals AI | 90.3% | 20.2%† |
| SWE-Bench Pro (private) | Scale AI SEAL | – | 34.6%† |
| SWE Atlas: Codebase QnA | Scale AI SEAL | 39.0%† | 8.2%† |
| SWE Atlas: Refactoring | Scale AI SEAL | 54.8%† | 10.0%† |
| SWE Atlas: Test Writing | Scale AI SEAL | 55.6%† | 30.3%† |
| MirrorCode | Epoch AI Benchmarking Hub | 63.9%† | – |
| Surface Evolver Bench | Epoch AI Benchmarking Hub | 95.0%† | – |
| DeepSWE v1.1 | Epoch AI Benchmarking Hub | 69.7%† | – |
| ALE-Bench | Epoch AI Benchmarking Hub | 2041.3† | 1367.2† |
| FrontierSWE | Epoch AI Benchmarking Hub | 47.0%† | – |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 42.4%† | – |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | 84.6%† | – |
| Startrise Frontend | Startrise | 73.8% | – |
| LiveCodeBench (AA) | Artificial Analysis | – | 90.8% |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 44.5%† | 64.3%† |
| Remote Labor Index | Scale AI / CAIS | 16.1% | – |
| APEX-Agents | Mercor | 63.6% | – |
| Arena Agent | Arena (LMArena) | 8† | – |
| LiveBench Agentic Coding | LiveBench | 62.2%† | – |
| Terminal-Bench Hard | Artificial Analysis | 62.9%† | 38.6% |
| τ²-Bench Telecom (AA) | Artificial Analysis | 98.5%† | 80.4% |
| Terminal-Bench 2.1 (Vals) | Vals AI | 80.5% | 53.9%† |
| MCP Atlas | Scale AI SEAL | 83.3% | 62.0%† |
| HiL-Bench | Scale AI SEAL | 56.3% | – |
| τ²-bench | τ²-bench | – | 83.5%† |
| τ³-bench Banking | τ²-bench | 39.7%† | 27.3%† |
| τ³-Banking (AA) | Artificial Analysis | 38.1%† | 20.8% |
| Analyst Agent (AA) | Artificial Analysis | 48.8%† | – |
| APEX-Agents (AA) | Artificial Analysis | – | 27.7% |
| Finance Agent v2 | Vals AI | 56.3% | 42.5%† |
| Harvey's Legal Agent Benchmark v1 | Vals AI | 11.3% | 0.0%† |
| Legal Research Bench | Vals AI | 49.5% | 18.3%† |
| Poker Agent | Vals AI | – | 1100.2%† |
| Tax Agent Bench | Vals AI | 38.3% | 11.8%† |
| Terminal-Bench 2.0 (Vals) | Vals AI | – | 51.7%† |
| Terminal-Bench 4.0 (Vals) | Vals AI | 41.4% | – |
| Terminal-Bench Science | Vals AI | 15.7% | – |
| Web Search Index | Vals AI | 46.9% | – |
| Arena Search | Arena (LMArena) | 1229† | – |
| BALROG | BALROG | – | 48.1%† |
| DeepResearch Bench | DeepResearch Bench | – | 47.9%† |
| PostTrainBench | Epoch AI Benchmarking Hub | 41.8%† | – |
| GBAEval | Epoch AI Benchmarking Hub | 74.5% | – |
| Vending-Bench 2 | Andon Labs | 4529.9 | 3634.7† |
| AutomationBench | Artificial Analysis | 54.1%† | – |
| GDP.pdf (AA) | Artificial Analysis | 24.0%† | – |
| EnterpriseOps-Gym | Artificial Analysis | 51.1%† | – |
| AA-Briefcase v1.1 | Artificial Analysis | 1535† | – |
| BrowseComp | OpenAI | – | 16.9%† |
| FACTS Search | – | 80.9%† | |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 87.0%† | 51.2%† |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 90.2%† | 17.1%† |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 99.7%† | 92.8%† |
| ProofBench | Vals AI | 95.0%† | 15.0%† |
| LiveBench Mathematics | LiveBench | 96.0%† | – |
| AIME (Vals) | Vals AI | – | 95.6%† |
| MGSM | Vals AI | – | 93.3%† |
| AIME 2026 | MathArena | – | 96.7%† |
| HMMT February 2026 | MathArena | – | 89.4%† |
| MathArena Apex | MathArena | – | 15.6%† |
| Arena Maths | Arena (LMArena) | 1521† | 1475† |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 70.7%† | 66.8%† |
| LiveBench 1.0 Data Analysis | LiveBench | 80.5%† | – |
| MMLU-Pro | Vals AI | 91.5% | 88.6%† |
| LegalBench | Vals AI | 88.6% | 86.9%† |
| CorpFin | Vals AI | 71.8% | 66.4%† |
| TaxEval | Vals AI | 76.9% | 73.9%† |
| MedQA | Vals AI | – | 95.8%† |
| PRBench Finance | Scale AI SEAL | 53.9% | – |
| PRBench Legal | Scale AI SEAL | 52.6% | – |
| CaseLaw v2 | Vals AI | – | 55.8%† |
| Excel Modeling Benchmark | Vals AI | 73.7% | 26.0%† |
| MedCode | Vals AI | 56.1% | 55.9%† |
| MedScribe | Vals AI | 88.5% | 69.9%† |
| Public Benefits Bench v1.1 | Vals AI | 70.4% | – |
| EBR-bench | Epoch AI Benchmarking Hub | 39.5%† | – |
| GDP.pdf | Epoch AI Benchmarking Hub | 30.0% | 10.0%† |
| Global-MMLU-Lite (AA) | Artificial Analysis | – | 92.8%† |
| Instruction following | |||
| LiveBench Language | LiveBench | 90.7%† | – |
| LiveBench Instruction Following | LiveBench | 75.8%† | – |
| IFBench | Artificial Analysis | 63.5%† | 78.0% |
| Arena Instruction Following | Arena (LMArena) | 1509† | 1458† |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1504† | 1473† |
| EQ-Bench 4 | EQ-Bench | 1313 | – |
| Arena Creative Writing | Arena (LMArena) | 1501† | 1457† |
| Arena Multi-turn | Arena (LMArena) | 1516† | 1482† |
| Arena Longer Queries | Arena (LMArena) | 1522† | 1476† |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1325† | 1285† |
| MMMU-Pro | Artificial Analysis | – | 79.9% |
| MMMU-Pro (Vals) | Vals AI | 89.3% | 87.6%† |
| MortgageTax | Vals AI | 68.9% | 68.7%† |
| GeoBench | GeoBench | – | 88.0%† |
| VPCT | Epoch AI Benchmarking Hub | – | 72.6%† |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 38.6% | 0.0%† |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 82.3%† | 78.0% |
| Arena Document | Arena (LMArena) | 1499 | 1439† |
| MLCR | Artificial Analysis | 20.6%† | – |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 163.1 | 151.9† |
| LiveBench 1.0 | LiveBench | 83.0%† | – |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 49.6† | 26.3 |
| AA Progress Index v1 | Artificial Analysis | 329.3† | 290.1 |
| Vals RSI Index | Vals AI | 24.3%† | – |
| Vals Multimodal Index | Vals AI | 74.2% | 52.2%† |
Data as of 2026-10-29. Best configuration of each model; every score links to its source on the model pages.
What changed
- Gemini 3 Flash (thinking): 1 new result (LiveCodeBench (AA)).
- Gemini 3 Flash (thinking): BenchLeader Index 60.1 → 59.0.
- Gemini 3 Flash (thinking): 1 new result (AA Progress Index v1).
- Claude Fable 5: BenchLeader Index 67.3 → 66.5.
- Claude Fable 5: 1 score revision.
- Claude Fable 5: 1 new result (Harvey's Legal Agent Benchmark v1).
- Claude Fable 5: 1 score revision.
- Claude Fable 5: 8 new results (LiveBench 2.0, LiveBench Long Agentic, LiveBench 2.0 Reasoning…).
- Gemini 3 Flash (thinking): 2 new results (AA-Omniscience v1.1: accuracy, AA-Omniscience v1.1: non-hallucination).
- Claude Fable 5: 1 new result (Startrise Frontend); 1 score revision.
Related comparisons
- Claude Fable 5 vs Claude Opus 5.5
- Claude Fable 5 vs Gemini 4 Carbon
- Claude Fable 5 vs GPT-6 Astra
- Gemini 3 Flash vs Claude Opus 5.5
- Gemini 3 Flash vs Gemini 4 Carbon
- Gemini 3 Flash vs GPT-6 Astra
- Gemini 3 Flash vs GPT-6.1 Sol
- Claude Fable 5 vs Gemini 3 Pro
Rankings: Best LLM overall · Best LLM for coding · Best LLM for agents · Best cheap LLMs
Claude Fable 5 vs Gemini 3 Flash: questions
- Is Claude Fable 5 better than Gemini 3 Flash?
- Claude Fable 5 leads on quality: 66.6 vs 59.0. The BenchLeader Index combines every independent quality benchmark; Claude Fable 5 is ahead overall as of 29 Oct 2026, but check the category scores for your use.
- Claude Fable 5 vs Gemini 3 Flash: which is cheaper, and will that change?
- Gemini 3 Flash is cheaper today: $1.13 against $20.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Which is faster, Claude Fable 5 or Gemini 3 Flash?
- Gemini 3 Flash streams 3×+ faster (209 vs 53 output tokens per second, Artificial Analysis 7-day median).
- Is Claude Fable 5 or Gemini 3 Flash better for agents and tool use?
- Claude Fable 5 scores higher in agentic tasks (64 vs 55 on the category index, where 50 is average).
- Which is better for long documents?
- Both take about 1M tokens: Claude Fable 5 1,000,000, Gemini 3 Flash 1,048,576.
More questions (2)
- Which has lower latency?
- Claude Fable 5 answers first: 3.8 s against 6.0 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
- When was Claude Fable 5 released?
- Claude Fable 5 was released on 9 Jun 2026.