Quick Take
- Anthropic launched Claude Opus 5 on July 24, 2026, priced at $5 (Rs 483) per million input tokens.
- Opus 5 tops coding and knowledge work benchmarks at half the cost of Fable 5.
- An effort setting lets teams trade intelligence for speed, cutting token spend on routine tasks.
In This Article
Opus 5 beats Fable 5 on coding and knowledge work benchmarks while costing half as much, Anthropic said on July 24, 2026. The new model is priced at $5 (Rs 483) per million input tokens and $25 (Rs 2,414) per million output tokens.
The claim needs one caveat. Opus 5 is not the smartest Claude model overall, that title still belongs to Fable 5 and Mythos 5. But on software engineering and professional tasks, Opus 5 sets a new state-of-the-art, per Anthropic’s launch announcement. It is now the default model on Claude Max and the strongest option on Claude Pro.
StartupFeed Insight
The real story is not raw power, it is unit economics. Anthropic is betting most valuable AI work sits in a middle band of difficulty, where near-frontier intelligence delivered cheaply beats a costlier top model. Enterprise buyers burned by Fable 5’s high token burn will notice. Watch Indian SaaS teams and AI startups, who run tight compute budgets, migrate coding and agent workloads to Opus 5 first. StartupFeed expects Anthropic to ship a Sonnet-tier refresh at this same price-for-performance pitch before the end of 2026, as it races OpenAI ahead of a planned IPO. By StartupFeed Desk.
Opus 5 Beats Fable 5: The Numbers
Opus 5 is Anthropic’s upgrade to the Opus tier, replacing Opus 4.8 at the same price. The table below sets out the core launch facts.
| Metric | Detail | Notes |
|---|---|---|
| Model | Claude Opus 5 | Upgrades Opus 4.8 |
| Launch date | July 24, 2026 | Available on all platforms |
| Input price | $5 (Rs 483) per million tokens | Same as Opus 4.8 |
| Output price | $25 (Rs 2,414) per million tokens | Same as Opus 4.8 |
| Cost vs Fable 5 | Half the price | On coding tasks |
| Frontier-Bench v0.1 | Doubles Opus 4.8 score | Tops all other models |
The standout figure: on Frontier-Bench v0.1, Opus 5 more than doubles Opus 4.8’s score at a lower cost per task, Anthropic reported. On CursorBench 3.2 at max effort, it lands within 0.5% of Fable 5’s peak, at half the cost.
About Anthropic
Anthropic is an AI safety company founded in 2021 by former OpenAI staff including Dario and Daniela Amodei, headquartered in San Francisco. It builds the Claude family of models across five tiers: Haiku, Sonnet, Opus, Fable, and Mythos. The company reached a roughly $350 Bn valuation in late 2025 and filed confidential IPO paperwork in June 2026.
Why is Opus 5 cheaper than Fable 5?
Opus 5 is cheaper because it targets everyday work, not the hardest autonomous tasks. Anthropic priced it identically to Opus 4.8 while nearly matching Fable 5’s intelligence on coding. A new effort setting, low, medium, or high, lets customers dial how many tokens the model spends, trading capability for cost.
“Claude Opus 5 delivers near Fable 5 intelligence at Opus speed and cost. On CursorBench it’s just under Fable 5 and has many of the same behaviors,” said Sualeh Asif, Co-Founder, Cursor.
Fable 5, released in June 2026, drew criticism for a high token burn rate that ran up large bills. Opus 5 answers that directly. Anthropic still recommends Fable 5 for advanced projects that run autonomously for days.
How does Opus 5 compare to rivals?
Opus 5 arrives in a crowded race where price now matters as much as peak scores. OpenAI shipped its GPT-5.6 family earlier in July 2026, and China’s Moonshot AI released the open-weight Kimi K3 model, which the lab said nears both Fable 5 and GPT-5.6 in capability.
| Model | Maker | Positioning |
|---|---|---|
| Opus 5 | Anthropic | Near-frontier coding, half the cost |
| Fable 5 | Anthropic | Top intelligence, higher token burn |
| Kimi K3 | Moonshot AI | Open-weight, low cost |
What sets Opus 5 apart is the cost-per-result framing. On several benchmarks it beats every rival at any given price point, not just at peak spend.
What Opus 5 means for Indian founders
For India’s developer and startup market, cheaper near-frontier intelligence lowers the cost of shipping AI features. Teams building coding agents, automation tools, and analytics products can now run heavier workloads without the Fable 5 bill. Anthropic also added automatic fallbacks, so flagged requests route to another model instead of failing. Will price-led competition push Indian AI startups to build on Opus 5 by default?
Frequently Asked Questions
Written by Avinash. Have a tip? Write to us at editorial@startupfeed.in.
