Claude Fable 5.2
aka Anthropic Claude Fable 5.2
AnthropicCoding model76 posts
Posts
76incl. 3 solo runsMiami as Vice City: Claude Fable 5.2 vs GPT-6.2 Astra: Claude Fable 5.2, GPT-6.2 Astra. Posted by @ArtificialAnlys.
Paper crane, night city, three models: GPT-6.2 Astra, Claude Fable 5.2, Gemini 4 Carbon. Posted by @HyunSeob_.
Pelican 3D: Claude Fable 5.2 vs GPT-6.2 Astra: Claude Fable 5.2, GPT-6.2 Astra. Posted by @GKev1n.
Pixel flying game: Claude Fable 5.2 vs GPT-6.2 Astra: Claude Fable 5.2, GPT-6.2 Astra. Posted by @Densitylucky.
Alpine Gold plane, four models: GPT-6.2 Astra, GPT-6.2 Sol, Claude Fable 5.2, Claude Opus 5.5. Posted by @EnvolDev.
Kerr black hole: GPT-6.2 Astra vs Claude Fable 5.2: GPT-6.2 Astra, Claude Fable 5.2. Posted by @filicroval.
Peacock SVG: GPT-6.2 Astra vs Claude Fable 5.2: GPT-6.2 Astra, Claude Fable 5.2. Posted by @R2Cdev_.
Alpine Gold plane, Qwen4-Max and three US models: Qwen4-Max, Claude Fable 5.2, Gemini 4 Argon, GPT-6 Astra. Posted by @EnvolDev.
Crown of the Reef, four models: Claude Fable 5.2, Claude Opus 5.5, GPT-6 Astra, GPT-6.1 Sol. Posted by @EnvolDev.
Pelican on a zebra on a horse, four models: Qwen4-Max, Claude Fable 5.2, GPT-6 Astra, Gemini 4 Hydrogen. Posted by @RezAIT00.
Walk training, three models: Claude Fable 5.2, Qwen4-Max, Gemini 4 Hydrogen. Posted by @noclipepe.
Pelican cycling in the rain, with splash and controls: Qwen4-Max, Claude Fable 5.2, Gemini 4 Hydrogen. Posted by @hiwise_aigc.
Rio above the clouds, one shot: Claude Opus 5.5, Claude Fable 5.2, Gemini 4 Hydrogen, GPT-6 Astra. Posted by @MartinSzerment.
Rio at night, above the clouds: Claude Fable 5.2, Claude Opus 5.5, GPT-6 Astra, Gemini 4 Hydrogen. Posted by @SamgreyAi.
Rio above the clouds: Claude Opus 5.5, Claude Fable 5.2, Gemini 4 Hydrogen, GPT-6 Astra. Posted by @EnvolDev.
The Crown of the Reef: Claude Fable 5.2, Claude Opus 5.5. Posted by @EnvolDev.
Spider-Man model in Three.js: GPT-6 Astra, Claude Opus 5.5, Claude Fable 5.2, Gemini 4 Argon, Gemini 4 Carbon, Gemini 4 Hydrogen. Posted by @VibeDevarf.
Avalanche simulation, a large cost gap: Claude Fable 5.2, Gemini 4 Carbon. Posted by @AIMind_Ai.
Night Tokyo 04, a rainy ramen stall: Claude Fable 5.2, GPT-6 Astra, Gemini 4 Hydrogen, Gemini 4 Argon. Posted by @sprinntt.
The Up house, same prompt and reference: Claude Fable 5.2, Claude Opus 5.5, Gemini 4 Carbon, Gemini 4 Hydrogen. Posted by @noclipepe.
Dragon leaping from The Great Wave: Claude Fable 5.2, Gemini 4 Carbon. Posted by @0xRhinoX.
Pirate ship in Blender: Claude Opus 5.5, Claude Fable 5.2. Posted by @MrTacticalX.
Arcade game, four models: Claude Fable 5.2, Gemini 4 Carbon, Claude Opus 5.5, GPT-6 Astra. Posted by @0xTeddet.
Claude Fable 5.2 vs GPT-6.2 Sol vs Gemini 4 Hydrogen: Claude Fable 5.2, GPT-6.2 Sol, Gemini 4 Hydrogen. Posted by @Stargridcompass.
NYC night walk into a laundromat, in Blender: Gemini 4 Hydrogen, Claude Fable 5.2. Posted by @colerepo.
Voxel US map, two Claude Fable versions: Claude Fable 5.1, Claude Fable 5.2. Posted by @minebench_ai.
Brand for a repair café that does not exist yet: Claude Fable 5.2, Claude Opus 5.5. Posted by @AsumaCodes.
Metal bench with a shoe rack, built for real: Claude Fable 5.2, Claude Opus 5.5. Posted by @emilylai.
Boat in a storm at sea, with a Blender run: Claude Fable 5.2, GPT-6 Astra. Posted by @m1lo_exe.
3D horse simulation and per-token prices: GPT-6 Astra, Claude Fable 5.2, Gemini 4 Carbon, Gemini 4 Hydrogen. Posted by @Flusku.
AI bill calculator with an over-the-top receipt: GPT-6 Astra, Claude Fable 5.2, Claude Opus 5.5, DeepSeek V4.1 Flash 1017. Posted by @kaireceipts.
Kyoto city scene, two Claude models: Claude Fable 5.2, Claude Opus 5.5. Posted by @joseph_mugen.
Animated window-garden scene, Claude Code vs Codex: Claude Fable 5.2, GPT-6 Astra. Posted by @sebmoska.
Flipping iceberg, about 30x cost gap: Claude Fable 5.2, Gemini 4 Carbon. Posted by @AIMind_Ai.
Chai seller platform game from one sentence: Claude Fable 5.2, Claude Opus 5.5. Posted by @AsumaCodes.
Eiffel Tower, aquarium and chonky cat test prompts: Claude Fable 5.2, Gemini 4 Carbon. Posted by @domsilvestre.
Souls-like boss fight in the browser, cross-post: Gemini 4 Carbon, Claude Fable 5.2. Posted by @ggff0i76.
Cow backflipping on a trampoline in Three.js: Claude Fable 5.2, Claude Opus 5.5, GPT-6.1 Sol, Gemini 4 Argon. Posted by @mmmarcus_ai.
Cobra bursting up through a palace carpet: Claude Fable 5.2, Gemini 4 Argon. Posted by @kinderlayer.
Flaming torch with sound design, high effort: GPT-6 Astra, Claude Fable 5.2. Posted by @camperaigg.
Eiffel Tower from one prompt, two Fable versions: Claude Fable 5.2, Claude Fable 5.1. Posted by @AIMind_Ai.
A trip planner you can talk to: Claude Fable 5.2, Claude Opus 5.5. Posted by @AsumaCodes.
Self-playing asteroid shooter in one HTML file: Claude Opus 5.5, Claude Fable 5.2. Posted by @Awanioz.
Pelican riding a bicycle, two Fable generations: Claude Fable 5.1, Claude Fable 5.2. Posted by @xmglab.
One-page website built from a launch article: Claude Fable 5.2, Claude Opus 5.5. Posted by @makaneko_AI.
Rocket launch from pad to orbit with real physics, solo run: Claude Fable 5.2 (solo run). Posted by @Shannnu_18.
Pomeranian, puddle, doghouse and bone: 12 s animation: Claude Fable 5.2, Claude Fable 5.1, Claude Opus 5.5, GPT-6.1 Terra. Posted by @Patryk70847138.
About Claude Fable 5.2
Every Claude Fable 5.2 comparison on aivs.tv: 73 side-by-side posts from X against other models on the same prompt, plus 3 solo runs on classic topics, 36 with cost or time reported. Compare coding ability, cost and speed from real tests.
Record against each opponent
Claude Opus 5.5 is cheaper than Claude Fable 5.2 in 14 of 16 posts that report costs for both.
| Opponent | Posts | Cheaper | Faster | Author picks |
|---|---|---|---|---|
| Claude Opus 5.5 | 40 | 2/16 | 11/21 | 4–0 |
| GPT-6 Astra | 16 | 3/10 | 2/9 | 1–0 |
| Gemini 4 Hydrogen | 12 | 0/6 | 1/6 | 1–0 |
| Gemini 4 Carbon | 10 | 0/5 | 1/3 | 1–1 |
| Claude Fable 5.1 | 7 | 1/2 | 1/3 | 2–0 |
| Gemini 4 Argon | 5 | 0/1 | 0/2 | — |
| GPT-6.1 Sol | 3 | — | 0/1 | — |
| DeepSeek V4.1 Flash 1017 | 1 | 0/1 | — | — |
Summary
Claude Fable 5.2 (Anthropic) appears in 76 posts on aivs.tv, next to GPT-6.2 Astra, Gemini 4 Carbon, GPT-6.2 Sol, Claude Opus 5.5, Qwen4-Max, Gemini 4 Argon and 8 more. Tasks: rocket launch: Claude Fable 5.2 vs GPT-6.2 Astra, miami as Vice City: Claude Fable 5.2 vs GPT-6.2 Astra, paper crane, night city, three models, pelican 3D: Claude Fable 5.2 vs GPT-6.2 Astra, pixel flying game: Claude Fable 5.2 vs GPT-6.2 Astra, alpine Gold plane, four models and 70 more. Reported for Claude Fable 5.2: $2.76 / 11.9 min in @MartinSzerment's rio above the clouds, one shot, $42.10 / 1 h 13 min in @SamgreyAi's rio at night, above the clouds, $76.04 / 1 h 59 min in @EnvolDev's rio above the clouds, $78.45 / 3 h 15 min in @EnvolDev's Crown of the Reef, $57.00 / 1 h 30 min in @AIMind_Ai's avalanche simulation, a large cost gap, $7.11 / 20 min in @sprinntt's night Tokyo 04, a rainy ramen stall, $19.10 / 45.7 min in @noclipepe's Up house, same prompt and reference, $34.41 / 1 h 33 min in @MrTacticalX's pirate ship in Blender, $22.00 / 9 min in @0xTeddet's arcade game, four models, $15.94 / 38.5 min in @colerepo's NYC night walk into a laundromat, in Blender, $209.50 / 3 h 03 min in @m1lo_exe's boat in a storm at sea, with a Blender run, $7.72 in @kaireceipts's AI bill calculator with an over-the-top receipt, $4.00 / 11.4 min in @sebmoska's animated window-garden scene, Claude Code vs Codex, $10.00 in @AIMind_Ai's flipping iceberg, about 30x cost gap, $17.00 in @domsilvestre's Eiffel Tower, aquarium and chonky cat test prompts, 13.2 min in @mmmarcus_ai's cow backflipping on a trampoline in Three.js, $21.30 / 1 h 25 min in @camperaigg's flaming torch with sound design, high effort, $4.88 / 20 min in @AIMind_Ai's Eiffel Tower from one prompt, two Fable versions, 5 min in @Awanioz's self-playing asteroid shooter in one HTML file, 6 min in @makaneko_AI's one-page website built from a launch article, 59.1K tokens in @Patryk70847138's pomeranian, puddle, doghouse and bone: 12 s animation, 50 min in @shirotapandasan's 3D miniature garden built from a reference image, $2.57 / 24 min in @Ochena250312's 3D flying game, one try each, $57.87 / 1 h 30 min in @MrTacticalX's glass rooftop pool from sunset to night, $6.97 / 21 min in @MrTacticalX's snow globe with a log cabin and a snowman, $0.24 / 60 s in @MervinPraison's pelican on a bike: effort settings compared, $23.45 / 32 min in @FabianoFirmo's rocket launch scene, $1.92 / 8 min in @MartinSzerment's honeybee landing on a coneflower, macro at golden hour, $6.86 / 19.5 min in @darrxgh's snow globe you can shake and grab, 4 min / 80K tokens in @itsarcvale's operating system that boots in the browser, $81.10 / 1 h 57 min in @EnvolDev's cathedral of Ice: a ship sailing among giant icebergs, $6.48 / 30 min in @Rockstox's pelican juggling bowling pins on a unicycle, $19.00 / 55 min in @realNuun's six one-shot games judged on look, rules and feel, $18.00 / 40 min in @vikktorrrre's voxel pagoda garden at the highest reasoning level and 10 min in @ITangieff's recreate a video of the Sun. @EnvolDev picks Claude Fable 5.2 as the best result in the Crown of the Reef. @noclipepe picks Claude Fable 5.2 as the best result in the Up house, same prompt and reference. @m1lo_exe picks Claude Fable 5.2 as the best result in boat in a storm at sea, with a Blender run. @KenGrayAI picks Claude Fable 5.2 as the best result in first-person kayak game through river canyons. @ArtificialAnlys picks Claude Fable 5.2 as the best result in two disk galaxies colliding and merging. @jan_volad picks Claude Fable 5.2 as the best result in building a profile picture in Minecraft. @vikktorrrre picks Claude Fable 5.2 as the best result in voxel pagoda garden at the highest reasoning level.
What each post reports for Claude Fable 5.2
Only numbers stated in each post. “—” means the post doesn’t say.
Who won where
9- The Crown of the Reef“Fable 5.2 wins again.”— the author’s view, @EnvolDev
- The Up house, same prompt and referenceThe poster says Fable built the best house, with the rest of the sentence cut off.— the author’s view, @noclipepe (paraphrased)
- Claude Fable 5.2 vs GPT-6.2 Sol vs Gemini 4 HydrogenClaude Fable 5.2 > GPT-6.2 Sol > Gemini 4 Hydrogen.— the author’s view, @Stargridcompass (paraphrased)
- Boat in a storm at sea, with a Blender runClaude Fable 5.2 wins with a stormy fishing trawler; GPT-6 Astra’s sunny pirate ship is solid for the price.— the author’s view, @m1lo_exe (paraphrased)
- Eiffel Tower from one prompt, two Fable versionsClaude Fable 5.2 is about 1.8× cheaper than Claude Fable 5.1 and twice as fast on this prompt.— the author’s view, @AIMind_Ai (paraphrased)
- First-person kayak game through river canyonsBoth performed well, but Fable is clearly more polished and its controls were much more responsive.— the author’s view, @KenGrayAI (paraphrased)
- Two disk galaxies colliding and mergingFable 5.2 produces a more detailed representation of the celestial collision.— the author’s view, @ArtificialAnlys (paraphrased)
- Building a profile picture in MinecraftFable is miles ahead; Opus came out almost primitive, though the poster notes it may be just this one test.— the author’s view, @jan_volad (paraphrased)
- Voxel pagoda garden at the highest reasoning levelFable 5.2 is a big jump: very little difference in cost and time, but a huge gulf in output quality.— the author’s view, @vikktorrrre (paraphrased)
Matchups with Claude Fable 5.2
14Spellings we map to Claude Fable 5.2
“claude-fable-5.2”, “Anthropic Claude Fable 5.2”















































