Claude Fable 5 vs Gemini 3.8 Flash
Vals Science Index?
Fable 5—
Gemini 3.8 Flash35.8
| Benchmark | Fable 5 | Gemini 3.8 Flash |
|---|---|---|
| Humanity's Last Exam | 55.5 | 42.1 |
| Arena Text · overall | 1,504 · high | 1,496 · high |
| Arena Text · Chinese | 1,542.3 · high | 1,550.8 · high |
| Arena Text · English | 1,499 · high | 1,499 · high |
| Arena Text · French | 1,503.6 · high | 1,491.7 · high |
| Arena Text · German | 1,486.7 · high | 1,497.6 · high |
| Arena Text · instruction following | 1,502.3 · high | 1,491 · high |
| Arena Text · Japanese | 1,506.4 · high | 1,506.3 · high |
| Arena Text · Korean | 1,491 · high | 1,461.6 · high |
| Arena Text · non-English | 1,480.7 · high | 1,491.1 · high |
| Arena Text · Polish | 1,493.4 · high | 1,498.5 · high |
| Arena Text · Russian | 1,503 · high | 1,514.4 · high |
| Arena Text · Spanish | 1,495.8 · high | 1,485.7 · high |
| AA Omniscience Index | 43.3 | 28.6 |
| SimpleQA Verified v1 | 70.7 · xhigh | 74.6 |
| CritPt (physics) | 28.6 | 12.3 |
| AA Intelligence Index 4.3.2 | 49.6 | 39.8 |
| AA Omniscience · Accuracy | 65.4 | 53 |
| AA Omniscience · Hallucination Rate | 63.6 | 51.9 |
| AA-LCR (long context) 1.1 | 82.3 | 84 |
| GDP.pdf · AA all-pass | 24 | 22.8 |
| MLCR-AA | 64.4 | 21.7 · high |
| ARC-AGI-1 · public evaluation | 97.6 | 97.8 |
| ARC-AGI-1 · semi-private | 98.5 | 97.5 |
| ARC-AGI-2 · public evaluation | 96.4 | 87.5 |
| ARC-AGI-2 · semi-private | 89.2 | 82.9 |
| Chess Puzzles | 47 | 58 · high |
| FrontierMath Erdos | 0 | 0 · high |
| FrontierMath Tier 4 (v2) | 90.2 | 22 · high |
| FrontierMath Tiers 1–3 (v2) | 87 | 68.4 · high |
| Mystery Game Puzzles | 52 | 47 · high |
| MathArena · ArXivMath · April 2026 | 72.5 | 64.2 |
| MathArena · ArXivMath · June 2026 | 85.4 | 67.4 |
| MathArena · ArXivMath · May 2026 | 87.5 | 58.3 |
| MathArena · BrokenArXiv · April 2026 | 54.1 | 33.6 |
| MathArena · BrokenArXiv · June 2026 | 47.8 | 20.8 |
| MathArena · BrokenArXiv · May 2026 | 44.5 | 19 |
| MathArena · ArXivLean · June 2026 | 15.2 | 19.6 |
| Vals LegalBench | 88.6 | 87 |
| Vals MedCode | 56.1 | 48.1 |
| Vals MedScribe | 88.5 | 84.5 |
| Vals Legal Research Bench | 49.5 | 38.9 |
| IFBench (instructions) | 63.5 | — |
| ARC-AGI-3 · semi-private | — | 4 |
| EQ-Bench 3 | 2,049.7 | — |
| MathArena · ArXivMath · August 2026 | — | 40.4 |
| MathArena · ArXivMath · September 2026 | — | 50 |
| MathArena · BrokenArXiv · August 2026 | — | 27.4 |
| MathArena · BrokenArXiv · September 2026 | — | 26.7 |
| Vals Science Index | — | 35.8 |
Specs
| Spec | Fable 5 | Gemini 3.8 Flash |
|---|---|---|
| Input, per 1M | $10 | !$0.75 |
| Output, per 1M | $50 | !$3.75 |
| Cost per task? | $8.75 | $0.93 |
| Speed, tok/s | 67 | — |
| Context | 1M | 1M |
| Weights | Closed | Closed |
| Released | Jun 2026 | Sep 2026 |
| See images | ✓ | ✓ |
| Read PDFs | Partial | ✓ |
| Hear audio | — | ✓ |
| Watch video | — | ✓ |
| Call tools | ✓ | ✓ |
| Force a tool call | Partial | Partial |
| Return strict JSON | ✓ | ✓ |
| Search the web | ? | ✓ |
| Run code | ? | ✓ |
| Use a computer | ? | ? |
What breaks when you switch
Fable 5 → Gemini 3.8 Flash
Sampling parameters are ignored›
temperature, top_p and top_k are ignored now and will cause HTTP 400 in future generations; remove them from requests.
No minimal thinking level›
The minimal thinking level returns an error; avoid minimal.
Gemini 3.8 Flash → Fable 5
Nothing known.