Fireworks Announces Major 30% Price Reduction on H200 141 GB GPU Hourly Rate, Now Just $6.99
What happened
Between the Apr 9, 2025 and Jul 6, 2025 captures of Fireworks's pricing page, Pulse recorded 9 changes: Developer plan tier removed from main pricing section (replaced with product-based presentation); Qwen3 30B model added at $0.15 input / $0.60 output per 1M tokens; H200 141 GB GPU hourly rate decreased from $9.99 to $6.99 (30% reduction); Fine-tuning pricing restructured: MoE tier pricing ($2.00-$6.00 per 1M tokens) removed and replaced with DeepSeek R1/V3 specific pricing at $10.00 per 1M training tokens; DeepSeek R1 0528 (Fast) model added at $3.00 input / $8.00 output per 1M tokens; Enterprise plan tier removed from main pricing section (replaced with product-based presentation); Qwen3 235B model added at $0.22 input / $0.88 output per 1M tokens; Yi Large model removed (was priced at $3.00 per 1M tokens); B200 180 GB GPU added at $11.99 per hour. Pulse files it as a pricing change (plan removed, feature added, price decrease, package change, feature removed) with a downward direction. Pulse's latest capture (Oct 19, 2026) lists On-Demand at $13 per GPU hour (B200 180GB) + $15 per GPU hour (B300 288GB), Serverless at $0.08 per 10M input tokens (embeddings, up to 150M params) + $0.16 per 10M input tokens (embeddings, 150M-350M params) and Fine Tuning at $0.5 per 1M training tokens, LoRA SFT (models up to 16B params).
Fireworks's previous recorded change was "Fireworks Unveils Major 10x Increase in Serverless Inference RPM for Developer Plan, Boosting Capacity to 6,000" (Q1 2025). Fireworks has been tracked since 2024-11-17.
In the same window Figma and Adobe also changed their Marketing & Content pricing pages. Those moves are listed further down, and the Pulse Agent can compare them side by side.
- Developer · PlanRemovedDeveloper plan tierNo Developer plan tier
- — · FeatureAddedNo Qwen3 30B modelQwen3 30B model
- — · Price−30%$9.99$6.99
- MoE · PackagingPackage change
Fine-tuning pricing restructured: MoE tier pricing ($2.00-$6.00 per 1M tokens) removed and replaced with DeepSeek R1/V3 specific pricing at $10.00 per 1M training tokens
- — · FeatureFeature addedNot listed$3.00
- Enterprise · PlanRemovedEnterprise plan tierNo Enterprise plan tier
- — · FeatureAddedNo Qwen3 235B modelQwen3 235B model
- — · FeatureFeature removed
Yi Large model removed (was priced at $3.00 per 1M tokens)
- — · FeatureAddedNo B200 180 GB GPUB200 180 GB GPU
| Plan | Field | Before | After | Change |
|---|---|---|---|---|
| Developer | Plan | Developer plan tier | No Developer plan tier | Removed |
| — | Feature | No Qwen3 30B model | Qwen3 30B model | Added |
| — | Price | $9.99 | $6.99 | −30% |
| MoE | Packaging | — | $10.00 | Package change |
| — | Feature | Not listed | $3.00 | Feature added |
| Enterprise | Plan | Enterprise plan tier | No Enterprise plan tier | Removed |
| — | Feature | No Qwen3 235B model | Qwen3 235B model | Added |
| — | Feature | — | $3.00 | Feature removed |
| — | Feature | No B200 180 GB GPU | B200 180 GB GPU | Added |
Values as written on Fireworks's public pricing page in the two captures Pulse compared; where the page states no number, the cells give the wording before and after.
Evidence
The two captures Pulse comparedFull-page captures of Fireworks's pricing page (fireworks.ai) from Apr 9, 2025 and Jul 6, 2025; the change is marked near the top of the page, where price reads $9.99 and then $6.99.
Sign in to see the evidence
The two full-page captures with this change marked up are free with an account.
Fireworks, the last 90 days
Full history- Week 3, 2026packagingFireworks Unifies Cached Pricing at 50% Across All Text and Vision Models
- Jan 10, 2026packagingFireworks Pricing Restructured: DeepSeek R1 0528 Sees 55% Input and 32.5% Output Price Reduction
- Q3 2025improvedFireworks Announces Major Price Decrease on H100 GPU, Now Just $4.00/hour (31% Off)
- Q2 2025packagingFireworks Announces Major 30% Price Reduction on H200 141 GB GPU Hourly Rate, Now Just $6.99 · this change
- Q1 2025improvedFireworks Unveils Major 10x Increase in Serverless Inference RPM for Developer Plan, Boosting Capacity to 6,000
Across Marketing & Content in the same window
All Marketing & Content changes- FigmaJan 11Figma Increases Professional Plan Price by 25% to $15/mo Amidst New Features and Plan Renaming
- AdobeJan 10Adobe Creativecloud Introduces Firefly Pro Plan at $29.99/mo with Early Access Pricing, Amidst Multiple Price Increases for Other Plans
- FigmaJan 11Figma Launches Figma Make (Beta) for AI-Powered Web App Development, Included in Full Seat Plans with Usage Limits
- FigmaJan 11Figma Restructures Pricing with New Seat Types, Monthly Options, and Significant Price Increases for Professional, Organization, and Enterprise Plans
- FigmaJan 11Figma Introduces AI Credits Allocation: Enterprise Plan Now Offers 4250 Credits/Month
Questions about this change
Did Fireworks raise prices in July 2025?
No — prices went down. Between Apr 9, 2025 and Jul 6, 2025 Fireworks's pricing page showed 1 price cut: $9.99 → $6.99 (−30%).
What changed on Fireworks's pricing page between Apr 9, 2025 and Jul 6, 2025?
Developer plan: Developer plan tier → No Developer plan tier, removed; feature: No Qwen3 30B model → Qwen3 30B model, added; price: $9.99 → $6.99 (−30%); Fine-tuning pricing restructured: MoE tier pricing ($2.00-$6.00 per 1M tokens) removed and replaced with DeepSeek R1/V3 specific pricing at $10.00 per 1M training tokens; feature: Not listed → $3.00; Enterprise plan: Enterprise plan tier → No Enterprise plan tier, removed; feature: No Qwen3 235B model → Qwen3 235B model, added; Yi Large model removed (was priced at $3.00 per 1M tokens); feature: No B200 180 GB GPU → B200 180 GB GPU, added. Pulse compared the captures it took on Apr 9, 2025 and Jul 6, 2025; the before/after table above lists each item as written on the page.
How often does Fireworks change its pricing?
Pulse recorded 1 other pricing-page change for Fireworks in the 90 days before this one; the previous change was "Fireworks Unveils Major 10x Increase in Serverless Inference RPM for Developer Plan, Boosting Capacity to 6,000" in Q1 2025; Fireworks has been tracked since 2024-11-17.